We are putting together genomic surveillance data provided by FISABIO-UV with mobility data from mobile networks to understand the propagation and risk of new variants. This effort is part of the PREPARE project under the [European COVID-X program for Data Solutions](https://www.covid-x.eu), which goal is to develop an automatic tool to inform in real time about the risk of these new variants according to mobility patterns and genomic detection.

We are testing with Omicron, which evolved in Spain from anecdotical to dominant in less than a month. We have reconstructed its evolution and how it expanded to the whole country.

Omicron had several entry points, and for the 1st of December it was potentially localized in Madrid, Barcelona, Bilbao, or Las Palmas de Gran Canaria, among other municipalities. We face a multi-seed scenario, which is the common situation along the pandemic (as we detected at the very beginning in 2020, not only in Spain but also in most of the European countries [1]).

According to the mobility patterns of the last weeks, this is how Omicron propagated (according to our empirical data and simulations) since 1st of December to today:

![1639926535772.gif](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400421/articles_pictures/1639926535772_zm0yla.gif)

With so many entry points and so distributed in the territory, its expansion looks unstoppable indeed. From this image, it is hard to see the underlaying patters and how each seed contributed to the total expansion. However, thanks to the digital simulations, we can isolate each seed to see its individual contribution for the locations with higher risk as for 1st of December: Madrid, Barcelona, Bilbao and Las Palmas de Gran Canaria.

This is how a new variant would expand if the focus is fully localized in Madrid:

![Madrid.gif](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400452/articles_pictures/Madrid_lkujea.gif)

We see how its central location plus the radial communication network generates a very efficient propagation to every corner of the country, arriving faster to regional capital cities as Barcelona, Zaragoza, Valencia, etc.

This is how it would expand starting in Barcelona:

![Barcelona.gif](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400408/articles_pictures/Barcelona_aat4qk.gif)

Its location at the North-East of the country mitigates the first stages of the expansion, making it more local and concentrated to the East coast. However, its large exchange of travellers with Madrid generates a second focus in the capital, expanding from there to the rest of the country.

Similar patterns are seen when the focus is Bilbao:

![Bilbao.gif](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400409/articles_pictures/Bilbao_lg3buj.gif)

The variant appears very fast both in Madrid and Barcelona, expanding to the rest of the country.

The effect is particularly clear when the entry point is in Las Palmas de Gran Canaria:

![Las Palmas.gif](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400446/articles_pictures/Las_Palmas_yieguf.gif)

The variant is already in Madrid even before it fully expands to the island, and from there diffuses to all the country.

We noticed early in the pandemic that Madrid has a central role in the expansion [2], and now we can visualize what we meant at that time. Madrid is hyper-connected with all other regions in Spain, and thus whatever happen in Madrid affects all the country. In addition, Spain is a tourism power hyper-connected with the rest of Europe [3], and thus what happens in Spain affects to whole continent. This is not an exception with the new variants!

Thanks David Mateo and the rest of Kido Dynamics team for your effort in making these results possible!

[1] [link](https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1009326)

[2] [link](https://www.medrxiv.org/content/10.1101/2020.05.09.20096339v2)

[3] [link](https://www.nature.com/articles/s41586-021-03677-y)



We are putting together genomic surveillance data provided by FISABIO-UV with mobility data from mobile networks to understand the propagation and risk of new variants. This effort is part of the PREPARE project under the ...


How Omicron expanded in Spain

“*Bombarding people with facts and exposing their individual ignorance is likely to backfire. People don’t like too many facts, and they certainly don’t like to feel stupid*”[1].

I will start with this sentence from Harari’s “*21 lessons for the 21st century*”. Most of the content I’m sharing in this article might be controversial from an industry point of view, but I consider necessary to clarify those. 

***Some Background info***

Location intelligence services has been a fast-growing sector over the last few years, mostly due to massive adoption of smartphones, and hence, digital advertisement exponential growth.

Most of the companies based their data in SDKs (software development kits). These SDK were initially conceived to enhance aspects such as functionality, connectivity or compatibility for a given APP or program. Since one the purpose of these SDKs was to gather individual information and device performance information to improve customer experience and/or APP performance, APP owners started to gather immense amount of individual, high valuable data. 

App publishers discovered a way to increase their revenues so, either they kept this data as proprietary (e.g. WhatsApp, Facebook) or opened their APPs to third parties, so they could incorporate new functionalities within their APPs, and monetize them. Some of these functionalities might include app monitoring, behavioral analytics, marketing automation, advertising, payments, and others. 

In this context, it’s easily understandable that most of the available location data comes from the ad tech ecosystem, being Bid Stream/Ad Exchanges the main source of this data. BS/AE provides publishers a set of opportunities to dynamically target potential customers when they open an APP or visit a webpage. This process is named Real Time Bidding (RTB), and it generates “bid requests”.

Each *bid request* has associated certain parameters, so the advertiser can decide to bid or not for that opportunity. These parameters might include: 

Timestamp
iOS Identifier for Advertisers (IDFA) or Google Advertising ID (AAID)
Location (LAT/LON)
And others

As you can imagine, this generates an important (and valuable) dataset, utilized not only by publishers and advertisers, but also by other companies providing location information. If you want to know how many SDKs are installed in one of your apps, you can find a guide in this [link](https://medium.com/@nadir/how-to-know-which-sdks-are-installed-on-a-mobile-app-for-free-a45380fb6be2). 

***Data Marketing***

Now we have a general idea about how data is gathered. This is not an exhaustive analysis, and there are other ways to access location data (e.g. a direct API from APP publishers), but it might be considered as a good proxy.

Everything that has to do with data, refers to the 4 Vs: volume, variety, velocity, and veracity. Location data companies utilize different proxies to dress their value proposition (yes, to dress, not to address), but the most common are:

- Active Devices/Users
- Mobile Apps

This information can be made readily available:

![Placer.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400449/articles_pictures/Placer_potahh.png)

Figure 1. Source: https://www.placer.ai

It can be hidden or masked with other statistics, so you must do the math:

![Olvin.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400447/articles_pictures/Olvin_ngwhax.png)

Figure 2. Source: https://www.olvin.com

![Near.co.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400442/articles_pictures/Near.co_tulwvt.png)

Figure 3. Source: https://near.co/data/

***Important to note:*** 5B events per day on 1.6B users is roughly 3 events per user per day. We will cover this point later.

Or directly not make accessible at all to prevent any unwanted query:

![Knowledge.near.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400426/articles_pictures/Knowledge.near_biqjop.png)

Figure 4. Source: https://knowledge.near.co

***Data crunching***

Some companies provide a more detailed data, which also requires some effort to properly interpret what it means. I’m going to do some math here:

![Doc.safegraph.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400411/articles_pictures/Doc.safegraph_dh0t9f.png)

Figure 5. Source: https://docs.safegraph.com/docs/patterns-summary-statistics

In this image we see median devices seen daily. For those no familiar with basic statistics, it basically means that putting in order how many devices are seen every day (day 1 - 100 devices, day 2 – 120, day – 130, etc.), you pick the 15th day. Is this a reliable proxy? As always, it depends. In normal distribution (Gauss’ bell), the median is equal to the mean[2], but in a power law[3], the mean is more evasive. Interestingly, human behavior’s related aspects such as population growth, income distribution or innovation’s diffusion follow power-laws. 

Let’s move next. Now we can find two new concepts: *monthly visits*, and *monthly visitors*.

![Doc.safegraph 2.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400410/articles_pictures/Doc.safegraph_2_ohdpcc.png)

Figure 6. Source: https://docs.safegraph.com/docs/patterns-summary-statistics

Bearing in mind previous numbers for July 2021 (» 16M *median devices seen*), having 38M monthly visitors (I assume here that one device represents one visitor), each device/visitor is visible just 16/38 = 42% of the days (on average). This is, a single user will be visible 12 days in each month, with no data for the other 18 days. So, 2 every 5 days.

Since this visibility index will most probably follow a power-law, we assume here (and will see after) that most of the devices are almost invisible for most of the month.

If we now pay attention very briefly to the visits, we can perform a quick calculation: 1’000/40 = 25 visits per visitor… *each month*. This already yields some conclusions: each person visits (*is visible in*) less than one place (POI) per day.

I raise here few questions: 

1. How many different places do you visit every month? 
2. And how many of those places do you repeat at least twice? Think about it. 
3. Would you consider this data as reliable?

***Data accuracy***

An additional component of location data USP is related to *GPS precision/accuracy*. I’m not going to extent myself too much on this point, since I prefer visual examples, but when we are talking about 30 meters precision[4], it represents a circle with 30 meters radius. That kind of accuracy is perfectly achievable in open areas with no building after few minutes. But we face here two issues:

1)    Most of the shops in densely populated urban areas are less than 30 meters apart from each other:

![GPS distance.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400424/articles_pictures/GPS_distance_qruop3.png)

Figure 7. Times Square (NY) - 7th Avenue has 30+ meters width

2)    GPS rarely works as expected in densely populated urban areas:

![GPS building.jpeg](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400424/articles_pictures/GPS_building_syzoqe.jpg)

Figure 8. Source: https://support.strava.com/hc/en-us/articles/216917707-Bad-GPS-Data

3)    And even St. Google creates weird scenarios:

![GPS Alberto.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400425/articles_pictures/GPS_Alberto_ybzlaf.png)

Figure 9. Screenshots realized by our CTO Alberto Hernando on a trip between Munich and Augsburg. Distance between Google's best guess and rail tracks is approximately 120 meters after 3 minutes

Even industry advocates such as Mobile Marketing Association acknowledges that “…*up to 60% of ad requests contain some form of location data. Of these requests, less than 1⁄3 is accurate within 50-100 meters of the stated location*”.[5] Note: this report is 5 years; if anyone finds a more updated analysis, I would really appreciate it. 

***Data clustering***

Finally, location data companies’ USP includes clustering as one the most important features of their data, based essentially on a pretty accurate individual information (but discontinuous as we have seen) and a GPS accurate precision (no comments). I agree that the ideal World when it comes to clustering would look like this:

![Unacast.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400469/articles_pictures/Unacast_sof7qc.png)

Figure 10. Source: www.unacast.com

Unfortunately, reality looks more like this:

![Hands.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400432/articles_pictures/Hands_ix9dyb.png)

Figure 11. Kido Dynamics based on SDK generated third party data

This plot is based on one month of data from 60M users in the largest city in Brazil. It’s important to bear in mind that for each of these points, it would be necessary to add a 10 meters radius buffer.

If you see any cluster or underlying pattern here bear in mind two aspects:

-       Even totally random data can gather in clusters[6]

-       Apophenia is a well-known cognitive bias[7]

Alternatively, some companies utilize different simplifications or aggregations to get a significant enough number of users (also known as “*critical mass*”). For that matter, they aggregate data within a certain geographic area such as in the example below:

![Placekey.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400452/articles_pictures/Placekey_dsfpeh.png)

Figure 12. Source: https://www.placekey.io

In this case, any point of interest (PoI) falling within one of those hexagons will have the exact number of unique visitors, visits, frequency, etc.

This approach is not good or bad, but misleading to a certain extent if data is shared based on specific PoI, as in this case:

![PoI.jpeg](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400451/articles_pictures/PoI_qdre9i.jpg)

***Data clarity and representativeness***

On a remarkable exercise of transparency, Quadrant (www.quadrant.io) shared till February 2020 a comprehensive analysis of what they call “*Data Quality Dashboard*”. It looks like this:

![Data Dashboard.jpeg](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400409/articles_pictures/Data_Dashboard_ejtodi.jpg)

For this month, they include 2 additional metrics: *daily average users* (DAU) and *monthly average users* (MAU). DAU equals 59.6M and MAU equals 400M, which gives a 14.9%.

If we pay attention to the graphics above, we can extract some very quick conclusions:

- 60% of the devices have less than 10 events per day, this is less than one event per hour
- 70% of the devices have <20 meters accuracy, which is pretty good
- 59% of the devices are visible just ***ONE DAY*** a month, and only 0.77% are visible everyday (this represents 3.08M of those 400M)
- 52% of the devices are visible just ***ONE HOUR*** a day, and only 0.79% are visible 24 hours a day

This example raises again the question about “*Data Representativeness*” or “*Sample Size*” that we discussed previously, and that some companies simply express as a percentage related to total population:

![Raw visits count.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400454/articles_pictures/Raw_visits_count_fjfrrd.png)

Figure 13. Source: https://shop.safegraph.com/

We encounter here a more complex and less straightforward question: since human behavior, internet activity, and hence, app visibility, all follow a power law (as plotted in the previous chart), without proper considerations of the size and scale limitations of such data, estimations of the population parameters are likely to be biased. Also, increasing sample size will not solve this problem, since power laws have scale invariance. 

***(Un) Timely data***

One final point regarding location intelligence is about “*reaching the right person, at the right time, in the right place*”.

We have seen so far that part of the value proposition is simply inexistant: the level of randomness associated to users’ visibility across multiple APPs in multiple locations during multiple days, it makes impossible to asseverate this principle of timeliness. 

Or not really: you might remember that most of digital ads companies are talking about CPM (cost per millage) and similar parameters when it comes to digital campaigns. Well, that’s the answer: if you can release a big enough number of advertisements, you can be sure that a significant number will hit your expected cohort. 

But this is not due to machine learning, deep learning, neural networks, or any other fancy word: it’s called “*law of large numbers*”. You remember why all these companies were worried about millions of devices and billions of events?

![Coincidence.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400402/articles_pictures/Coincidence_dufqve.png)

***Data conclusions***

So why such a secrecy and opacity regarding data accuracy, reliability, representativeness, etc? The answer is clear: data quality is incredibly poor, and the only way to somehow hide this reality is by fooling people with big numbers.

Google, Facebook, Amazon, have created an unnecessary obsession about micro-targeting and hyper-profiling, but they also play with cohorts and the law of large numbers. People are incredibly random in their behavior, and we can just provide a best guess or confidence margin when it comes to forecast what people will do. They are simply much better than others because they have much more data and to some extent more relevant, but this will not definitively apply to other APP publishers. 

“*Bombarding people with facts and exposing their individual ignorance is likely to backfire. People don’t like too many facts, and they certainly don’t like to feel stupid*”[8].

High resolution in space is useless without enough resolution in time. Why? Time gives the context, the motivations, and story behind why we go where we go. Being very precise only in location misses the insights that makes this information useful. Just like the famous mathematician joke:

*An engineer and a physicist are in a hot-air balloon.*

*After a few hours they lose track of where they are and descend to get directions.*

*They yell to a jogger, "Hey, can you tell us where we're at?"*

*After a few moments the jogger responds, "You're in a hot-air balloon."*

*The engineer says, "You must be a mathematician."*

*The jogger, shocked, responds, "yeah, how did you know I was a mathematician?"*

*"Because it took you far too long to come up with your answer, it was 100% correct, and it was completely useless." *
 
Even mathematicians will agree with us since there is a mathematic relationship between the precision in space and the precision in time to get meaningful results. Indeed, if we are interested in visits to certain locations and its frequency or popularity, the precision in how we define a location and the precision in the frequency of visits are related and limited by the resolution *in space* **AND** *in time*, like the Heisenberg’s uncertain relation. (We will discuss the mathematical grounds of this relation in a future technical paper.)

Today, the only source of data able to guarantee this equilibrium in the precision of both location and frequency is the mobile phone network signaling data. It sacrifices some resolution in space but offers high frequency data in time (50 to 500 events per day, on average). Being specific, mobile phone data makes 80% of the people “***visible***” at least 8 hours a day, and 70% are visible at least 16 hours, while 50% of the users are visible ALL DAYS in each month. 

![Comparaison.jpeg](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400404/articles_pictures/Comparaison_wyuzi5.jpg)

This richness (and broadness and depth) is essential when it comes to make sound decisions in aspects such as traffic modelling, retail expansion or tourism evaluation. It’s necessary to evaluate the quality of data we are utilizing before we make our decisions and being critical if we do not feel comfortable with that data. 

Next time you acquire a data set to improve your operations, performance, or strategy, be sure you don’t feel stupid afterwards.

-----------------------------------------------

This article has been written in collaboration with Dr. Alberto Hernando, CTO & Cofounder at Kido Dynamics

[1] Harari, Y. (2019). 21 lessons for the 21st century. ISBN 9781784708283

[2] [link](https://en.wikipedia.org/wiki/Normal_distribution)

[3] [link](https://en.wikipedia.org/wiki/Power_law)

[4] [link](https://www.placeiq.com/2016/06/mobile-location-data-is-accurate-up-to-30-meters-report/)

[5] [link](https://www.mmaglobal.com/files/documents/location-data-accuracy-v3.pdf)

[6] [link](https://www.r-bloggers.com/2018/11/generate-datasets-to-understand-some-clustering-algorithms-behavior/)

[7] [link](https://en.wikipedia.org/wiki/Apophenia)

[8] Harari, Y. (2019). 21 lessons for the 21st century. ISBN 9781784708283

The Naughty Bits of Location Intelligence Data

***The (tiny niche) opportunity?***

Again, lend from European Commission smart mobility strategy[1]: “*it is crucial that mobility is available and affordable for all, that rural and remote regions are better connected, accessible for persons with reduced mobility and persons with disabilities, and that the sector offers good social conditions, reskilling opportunities, and provides attractive jobs*”.

Industry players claim that MaaS’ target segment differs from traditional commuters, arguing that businesspeople or tourists will be willing to pay to a service that will guarantee a minimal friction in their tight schedules. 

In this context, most of existing MaaS solutions and providers are aiming to fulfill a very ***individual centric*** approach (which in essence clashes with previous statement): travel from wherever you are to wherever you want at whenever time you fancy. This encounters a simple but realistic issue: transportation infrastructure has a finite capacity, so not everybody can travel at the same time to the same or different place, unless we accept major by side consequences, such as increasing congestion[2] or excess capacity[3]. 

Demand studies, infrastructure planning, urban development, and other related fields, have spent centuries[4] trying to figure out how to provide the most meaningful, seamless, cost-efficient mix of transportation alternatives to citizens, optimizing the effectiveness of a community's land use and infrastructure. It’s necessary to analyze land use compatibility, as well as economic, environmental, and social trends. 

Any plan for any community should have into consideration aspects such as sustainability, pollution, existing infrastructure, congestion, crime, land values, economic development, social equity, zoning codes, and other legislation.

It’s clear that people are willing to switch to more sustainable transportation modes when certain conditions are met (cost, availability, convenience, schedule)[5]. Furthermore, MaaS defenders claim that citizens will make sound and reasoned decisions when it comes to choose between owning a car, share it or use it occasionally (sic[6]). 

But mobility is still expensive for low-income classes, and not sufficiently accessible for people with disabilities, reduced mobility, or low IT-literacy[7], without mentioning how mobility needs dramatically change over time linked to different life stages (e.g., kids, teenagers, adults, single, couple, couple with kids, elderly[8]…).

Meanwhile, people are far from being rational when it comes to make decisions[9], since “*people facing a risky choice leading to gains, individuals are risk-averse, preferring solutions that lead to a lower expected utility but with a higher certainty*”, as initial trials with Whim Helsinki and Whim West Midlands demonstrated with a low engagement ratio[10].

We can end up having a system of mobility’s castes, where those with financial capacity can afford expensive, all-in integrated solutions combining multiple transportation modes (taxi, train, ride hailing, e-scooter…), utilising public infrastructure as a private lane, and those with much more limited financial capacity but also with more specific unfulfilled transportation needs who are left behind[11].

***Anti-mobility***

It’s important to bear in mind that 50% of the population spend their daytime within less than 2.5 km (radius of gyration[12] < 2.5 km), a percentage that after COVID-19 confinement measures has increased to 65%. If we consider a 5.0 km radius, we have 70% and 75% of the population respectively.

![Radius of gyration.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400453/articles_pictures/Radius_of_gyration_o0ou5h.png)

In this context, MaaS has a specific target fixation with car owners. According to UK statistics[13], in England in 2018, 25% of trips were under 1 mile (1.6 km), and 68% under 5 miles (8.0 km), which is pretty much in line with our previous analysis. Of all car trips, 58% are performed within that range. In Switzerland[14], a country with a very efficient public transport system, 65% of all daily trips were performed by car, while 24% were performed on public transport; they also cover 36.8 km on average per person per day (!!!!). In both cases, no median data is provided regarding distance traveled, only average data (*if I eat one chicken and you don’t eat any, we have eaten half a chicken each*), which to large extent limit the depth of the picture.

It would be necessary in this context to understand who is massively utilising private cars and why, before demonising them. Most common non-technical reasons include safety (65%), suitability for everyday use (47%), driving comfort (47%), or spaciousness (37%)[15].

In my case, with two kids (3 and 6 years old), I might endorse all these reasons, even if I utilise our private car exclusively on the weekends. 

I would add here two additional psychological related aspects:

- Resistance to change or loss aversion, described previously for Whim, and reflecting on the influence that our first choice has in our subsequent decisions.
- Isolation, which might be more evasive, but many people consider his/her 45 minutes daily commuting in their car as their only daily peaceful island.

It’s also necessary to bear in mind how fast population is aging in most of European countries[16]. This will add new components to our urban environment, promoting the *15 minutes city vision* [17], where all necessary services should be within 15 minutes walking distance, including health centres, supermarkets, schools, etc. 

Then, there’s COVID-19, which has dramatically changed our lifestyle, boosting remote working, reducing our out of home consumption, encouraging on-line purchase, and reducing dramatically the distance we cover daily[18].

Should we allocate resources to solutions that are focused on perpetuating an existing mobility model but with a digital halo effect? Or should we rethink from scratch the way we have built cities and make full integrative decisions based on actual facts?

***Where is the data?***

“*Most of our views are shaped by communal groupthink rather than individual rationality, and we hold on to these views our of group loyalty. Bombarding people with facts and exposing their individual ignorance is likely to backfire. People don’t like too many facts, and they certainly don’t like to feel stupid* [19]”.

When people have to confront actual data, they become sensitive very quickly. When we analyze how many people is actually performing multimodal trips in any city or region, this number is dwarfed compare to the total number of trips actually performed[20].

In the particular case of the city of Madrid, just 14% of monitored trips are multimodal, and most of them already happen within the city center. This number is even lower in Barcelona (7.3%), Zaragoza (7.4%), Valencia (5.6%) and Bilbao (1.2%).

Now, if mobility has reduced even further with COVID-19, people are working remotely, and people are aging in specific neighbourhoods, what is our strategy? Because it’s clear that MaaS is solving a tiny, insignificant, negligible part of a multi-phazet problem, that we are still reluctant to tackle with actual data. 

Data is necessary to properly schedule and plan mobility services, including public transport, bike-sharing or free-floating vehicles, but also to analyse mobility patterns, identify existing and future needs, and prioritise actions. But this is the elephant in the room: who’s the first one to share its data? 

Public initiatives towards open data are remarkable, but (all) companies based their business in capturing, processing, and reselling private data to third parties, hence making publicly available this data would undermine their competitive advantage. 

In that context, mobility data is not yet considered as a priority for public institutions, but as a good-to-have, so very little initiatives are driven to actually purchase this data from private players, with some exception during the recent COVID-19 pandemic. 

***Conclusion***

We need to figure out how we want to sketch the future of mobility. Things are changing dramatically, especially due to COVID-19. Remote working, social distance, safety travel, are all new paradigms that will shift our near and distant future, but we need to evaluate it, at least for now, and if we are just leveraging tech for the tech’s sake[21]. Companies, start-ups, investors, have all perceived an opportunity to intermediate an ecosystem that with certain inefficiencies has worked properly over the last decades, and it has allowed multiple societal layers to be integrated and get access to a fair transportation infrastructure. 

There are of course things to be changed an improved, but it would not be reasonable to allow (only) private players to take over a public service that has been developed very costly. 

I still remember how much easier my life became when Metro de Madrid opened *Line 12* in 2013[22]. It reduced my commuting time to the university by 15 minutes per leg (over an initial 1.5 hour) and limited how many times I had to change transport mode (bus to train to metro to bus, for metro to metro to bus). I didn’t have an app, everything was on a magnetic public transport ticket, and it was very cheap and convenient, since I was under 22.

I was a clear and obvious target for this service, as many other students by that time. We can now look for further efficiencies, better infrastructure, less congestion, but we will still need to answer a single question: how can we bring as many people as possible from point A to point B at the lowest possible cost?

This is something that MaaS cannot solve. At least not with the current rhetoric. It’s necessary to rethink how private companies want to interact with existing mass transport, and this interaction can bring some value to the end user. And this is necessary through a thorough analysis where transport authorities need to evaluate the impact of such players interfering with existing infrastructure, cost effectiveness, investment, target market, shifting costs, and most importantly, understanding what the actual needs of population are integrating all necessary data sources to make sound and informed decisions. 

I doubt all this will be possible with an app.

[1] [link](https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52020DC0789)

[2] [link](https://www.theverge.com/2019/8/6/20756945/uber-lyft-tnc-vmt-traffic-congestion-study-fehr-peers)

[3] [link](https://www.scmp.com/abacus/tech/article/3103485/abandoned-ofo-bikes-cry-out-help-while-piled-bike-sharing-graveyard)

[4] [link](https://en.wikipedia.org/wiki/History_of_urban_planning)

[5] Eurobarometer 495

[6] [link](https://www.gartner.com/en/podcasts/talent-angle/why-humans-arent-rational)

[7] [link](https://techcrunch.com/2019/05/05/we-are-leaving-older-adults-out-of-the-digital-world/)

[8] [link](http://www.internationaltransportforum.org/Pub/pdf/01Ageing.pdf))

[9] [link](https://en.wikipedia.org/wiki/Prospect_theory)

[10] Questioning mobility as a service: Unanticipated implications for society and governance, Pangbourne, Mladenovic, Stead, Milakis

[11] [link](https://news.gallup.com/poll/237965/snapshot-uses-ride-sharing-services.aspx)

[12] [link](https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0207697)

[13] [link](https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/823068/national-travel-survey-2018.pdf)

[14] [link](https://www.are.admin.ch/are/en/home/mobility/data/mtmc.html)

[15] [link](https://www.statista.com/chart/13075/most-important-factors-when-buying-a-car/)

[16] [link](https://ec.europa.eu/eurostat/statistics-explained/index.php/Population_structure_and_ageing)

[17] [link](https://citymonitor.ai/environment/what-is-a-15-minute-city)

[18] [link](https://www.medrxiv.org/content/10.1101/2020.10.26.20219709v1.full)

[19] Harari, Y. (2019). 21 lessons for the 21st century. ISBN 9781784708283

[20] Plan de Movilidad Urbana Sostenible de la ciudad de Madrid 2012, RESUMEN DIAGNOSTICO

[21] [link](https://www.bbc.com/news/technology-49674533)

[22] [link](https://en.wikipedia.org/wiki/Line_12_(Madrid_Metro))



The (Big) Fallacy of Mobility as a Service - Part 2

***Back to the future***

Imagine that you can pick a bus waiting less than 2 minutes at your bus stop, change to a train line with minimum lead time, travel 30 km, then change to a metro line for 4 stations, and pick a bike to ride your last 500 meters, and everything without buying a ticket. This is what you can do… in September 2000.

Actually, this is what I used to do when I was a student, with a small exception: I covered last 500 meters walking, and if I had a bike, I would need to lock and unlock it manually.

Now, let’s start with a simple question: can you spot 7 differences between next two Wikipedia entries:

**Mobility-as-a-Service (MaaS)** is a type of service that through a joint digital channel enables users to plan, book, and pay for multiple types of mobility services. The concept describes a shift away from personally-owned modes of transportation and towards mobility provided as a service. This is enabled by combining transportation services from public and private transportation providers through a unified gateway that creates and manages the trip, which users can pay for with a single account. Users can pay per trip or a monthly fee for a limited distance. The key concept behind MaaS is to offer travelers mobility solutions based on their travel needs[1]. 

The **Consorcio Regional de Transportes de Madrid** (CRTM; literally: *Regional Consortium of Transportation for Madrid*) is an autonomous body created by Spanish law 5/1985 which is tasked with coordinating the public transport operations across multiple providers in the Community of Madrid. It harmonises fares for commuter rail, rapid transit, light rail and bus transport services provided by entities such as Renfe Cercanías, Metro de Madrid S.A. or the Empresa Municipal de Transportes de Madrid (EMT)[2].

Actually, there’s just one difference: private transportation. Ok, you might consider “*joint digital channel*” as a key differentiator, but back in 2000 (and earlier), even if there were not digital platforms, you could still buy your “*abono transporte*” in any train or metro station. It looks like things are now digitizing. Good news [3].

I talk about Madrid since it’s the city I lived in and commuted to as a student and as an early graduate, but this might apply to many other cities Worldwide. For many years, I enjoyed a pretty efficient transportation system, with systemic break downs, delays, traffic jams, reschedules… you name it. But it worked. And even in extraordinary circumstances, the system has met it basics requirements[4].

What really freaks met out today is all hype happening around a concept which is more than 40 years old[5],[6],[7]. “*The future of transportation*”, “*How MaaS is disrupting mobility*”, “*The rise of Maas*”, …

Actually, what really really freaks me out is that companies and start-ups are racing into a market or sector that has been loss-making over the last 50 years[8],[9],[10],[11], and of course, venture capital money has followed short after[12].

According to the European Commission, “*the most serious challenge facing the transport sector is to significantly reduce its emissions and become more sustainable*[13]”, and “*digitalization will become an indispensable driver for the modernization of the entire system*”. It looks like to digitize something you just need to put it into an app, and then you can be worth billions. The difficulty comes to differentiate between a unicorn and a zebra – *a company providing an elegant, digital technology fuelled innovation when much simpler and cheaper solutions already exist*[14].

Question is if we are applying technology in the right order to fix our existing problems[15]. In my opinion, to some extent MaaS has been commercialized without being invented. And we all might face the consequences. 

***The business cases***

I’ve been in the VC industry, either as investor or as entrepreneur in the mobility market over the last 5 years. I have learnt the hard way how difficult is to put together a business, get funding, initial traction, new customers, etc. During that time, I have put my hands on some deck coming from MaaS start-up, and this is what you would need to do to justify a USD 523B market size[16]:

- In Madrid, there are 1’500M trips every year[17] utilizing some of the existing transportation modes (train, metro, bus…)
- This represents 230 annual trips for every of its 6.5M inhabitants
- 26.5% of the global population will live in urban agglomerations above 1M people by 2027[18]
- This represents 2.15B people
- These people will perform 230 trips annually on average utilizing a mix of public and private transportation modes  
- This represents 495B trips
- We will up-level including e-scooters, e-bikes, ride sharing… which will actually increase the willingness to pay and to utilize this platform among users, but we do not compute it
- Finally, we think we can capture up to USD 1.00 for every trip

(Note a caveat: all trips from everybody every time are booked through my app so I’m capturing all possible money available -called total addressable market or TAM-. Actually, less than 1% might be reasonable[19].)

Now you put all these data with some fancy charts, MBB style slides, include some testimonials, and voila, you can raise few millions from any VC.

For someone with certain experience in the transportation industry (I was Civil Engineer before becoming entrepreneur), there is an inherent flaw in this reasoning:

If you have an industry which is essentially loss-making and include some intermediation (even if digital), you are increasing the number of stakeholders, so you have more people at the table to share the same cake. Now, you are inviting even more guests, but from neighborhoods which are not precisely cash positive[20],[21]. So, if up to 50% of public transportation cost is already subsidized, how much money will be needed to have all these new guests (happy) at the table?

There’s a clear and intrinsic phenomenon around new business model built on top of what we used to knowing as public services: a taxi with an app is still a taxi, a delivery service with an app is still a delivery service, and a public transport ticket in an app is still a public transport ticket. ***Appstupidity***. 

If end customers maintain their level of willingness to pay and total expenditure for any of these services, someone will lose money in the value chain, either the driver, the rider, or whoever is providing a valuable piece in this chain.

In other industries, the process of value capture across value chain has been broadly analyzed, as well as some protective strategies against the so called “*value vampires*” (a disruptive player whose competitive advantage shrinks overall market size[22]). They systematically drain margin from incumbents, promoting “*customer empowerment, instant gratification and reduced friction*[23]”. Sounds familiar? 

The problem we encounter here is more elusive: incumbents are public services financed with people’s taxes. If they reduce their margins because value vampires intermediate some services, they will increase their deficit, and will need to look for new revenue sources to feed a larger stakeholders’ pool. This has traditionally led to two situations: taxes increase or privatization. 

In both cases, people will bear the consequences (cost), either with an increase of taxes and fares, or with higher prices driving to an expected profitability of private players, or with lay offs for redundancies. 

[1] [link](https://en.wikipedia.org/wiki/Mobility_as_a_service)

[2] [link](https://en.wikipedia.org/wiki/Consorcio_Regional_de_Transportes_de_Madrid)

[3] [link](https://pilotoapprecargattp.crtm.es/#item0)

[4] [link](https://elpais.com/espana/madrid/2021-01-10/que-abre-y-que-no-en-madrid-tras-el-temporal-filomena.html)

[5] [link](https://www2.deloitte.com/content/dam/Deloitte/nl/Documents/consumer-business/deloitte-nl-cb-ths-rise-of-mobility-as-a-service.pdf)

[6] [link](https://smartrak.com/maas-the-future-of-transportation/)

[7] [link](https://developer.ibm.com/recipes/tutorials/why-mobility-as-a-service-is-the-future-of-transportation/)

[8] [link](https://www.nytimes.com/2020/07/21/nyregion/mta-subway-financial-cuts.html)

[9] [link](https://www.bloomberg.com/news/articles/2020-12-22/transit-gets-a-reprieve-not-a-rescue-in-stimulus)

[10] [link](https://www.publico.es/economia/transportes-madrid-emt-tendra-perdidas-ano-frente-prevision-inicial-beneficios-equipo-carmena.html)

[11] [link](http://www.camaradecuentasmadrid.org/admin/uploads/aprobado-6.pdf)

[12] [link](https://tracxn.com/d/trending-themes/Startups-in-Mobility-as-a-Service)

[13] [link](https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52020DC0789)

[14] [link](https://www.imd.org/research-knowledge/articles/The-problem-with-digital/)

[15] [link](https://www.fastcompany.com/90664283/why-high-profile-smart-cities-fail-from-sidewalks-quayside-to-amazons-hq2-in-queens)

[16] [link](https://www.globenewswire.com/news-release/2020/12/14/2144252/0/en/Mobility-as-a-Service-Market-Size-Worth-USD-523-61-Billion-by-2027-Emergen-Research.html)

[17] [link](https://www.comunidad.madrid/inversion/madrid/transporte-publico-excelencia)

[18] [link](https://data.worldbank.org/indicator/EN.URB.MCTY.TL.ZS)

[19] [link](https://www.bloomberg.com/news/articles/2020-08-05/the-struggle-to-make-mobility-as-a-service-make-money)

[20] [link](https://www.theverge.com/2019/8/8/20793793/uber-5-billion-quarter-loss-profit-lyft-traffic-2019)

[21] [link](https://www.calcalistech.com/ctech/articles/0,7340,L-3764203,00.html)

[22] [link](https://www.imd.org/research-knowledge/reports/digitalvortex2019/)

[23] Digital Vortex: How Today's Market Leaders Can Beat Disruptive Competitors at Their Own Game, 2017, by Jeff Loucks, James Macaulay, Andy Noronha, Michael Wade

The Big Fallacy of Mobility as a Service (SaaS) - Part 1

Desde el inicio de la pandemia de COVID-19 en febrero de 2020, las distintas administraciones públicas españolas se han visto en la obligación de tomar diversas medidas con el fin de frenar la expansión del virus.

Además de las intervenciones sanitarias propiamente dichas (protocolos hospitalarios, medidas de higiene, etc.), las administraciones también tomaron otras de carácter no sanitario como los confinamientos, toques queda y cierres perimetrales. Sin ninguna duda, la más radical de estas medidas fue el confinamiento total de la población durante el primer Estado de Alarma entre el 15 de marzo y el 21 de junio de 2020[1].

Numerosos estudios científicos demuestran que el mayor factor en la expansión territorial de la enfermedad es la propia movilidad de las personas, tanto a nivel nacional como internacional. En particular, un reciente estudio publicado en la prestigiosa revista Nature[2] y participado por Kido Dynamics, demuestra como la movilidad nacional e internacional fue un factor decisivo en la expansión de nuevas variantes en el continente.

Con la eliminación del segundo Estado de Alarma el 9 de mayo de 2021[3], la movilidad de las personas ha vuelto a los altos valores pre-pandemia, y con dicha movilidad, se ha producido de manera paulatina un aumento de los contagios y por tanto de la incidencia acumulada en distintas regiones de España[4].

Los criterios establecidos por las administraciones públicas en estos casos se han centrado en dos aspectos fundamentales: la presión hospitalaria y la incidencia acumulada a 14 días (IA14) por cada 100’000 habitantes. En este sentido, la Junta de Andalucía estableció en el BOJA Extraordinario nº 41 de 07/05/2021, que una IA14 superior a 1’000 para aquellos municipios de más de 5’000 habitantes conllevaría el cierre perimetral de dicho municipio[5].

Lo que no dejaba claro el anuncio es la **población de referencia** que se debía utilizar para realizar dicho cálculo de la incidencia acumulada.

Por defecto, y de nuevo dentro del ámbito de las Administraciones Públicas, se considera que la población real de cada municipio corresponde a la población censada[6] o empadronada[7], es decir, al conjunto de aquellas personas que declaran su residencia en el municipio considerado.

Aunque desde un punto de vista administrativo este concepto es perfectamente correcto, desde un punto de vista práctico lleva aparejadas ciertas limitaciones: el censo no toma en cuenta la población flotante, esta es, el incremento de población que se da lugar durante el verano o los fines de semana, que en muchas ocasiones multiplica varias veces a la población oficial. 

Este hecho supuso recientemente un problema en la gestión sanitaria de varias Comunidades Autónomas, donde los servicios están escalados a la referencia de la población censada pero donde la población real aumentó al desplazarse muchos residentes externos a su segunda residencia dentro de la región. En particular, Castilla-La Mancha vio en 2020 cómo muchos desplazados desde Madrid hicieron incrementar notablemente la prestación de servicios sanitarios por encima de lo esperado[8],[9].

***Incidencia acumulada, población real, población censada y Big Data***

El municipio de Conil de la Frontera se ha enfrentado en los últimos días a una situación donde las cuestiones administrativas también han jugado claramente en su contra: con una población censada de unos 23’000 habitantes, el municipio había contabilizado en torno a 240 positivos durante los últimos 14 días, situando al municipio con una IA14 de 1’058. Este valor se situaba por encima del margen establecido por la Junta de Andalucía y acarreaba el cierre perimetral en plena temporada de verano, suponiendo un alto impacto social y económico para un municipio costero como Conil.

Sin embargo, es sabido que la población flotante de Conil de la Frontera en estas fechas es mucho mayor a lo que dice el censo, pues la ocupación hotelera y de segundas residencias hace que se concentre mucha mas gente en el municipio en la temporada de verano. Esto supone que la estimación del IA14 está sobreestimada. Pero ¿por cuánto exactamente? ¿Cómo podemos saber cuál es la población real de Conil de la Frontera en estas fechas?

Tanto el municipio de Conil de la Frontera como Kido Dynamics somos miembros de la Red de Destinos Turísticos Inteligentes promovida por Segittur, y mantenemos una estrecha colaboración desde hace tiempo.

En ese contexto, el Patronato de Turismo de Conil, representado por Diana Andión, tiene acceso diario a la información facilitada por Kido Dynamics a través de su plataforma de *smart tourism* https://destinosinteligentes.app, donde se recogen de manera pormenorizada los datos relativos a visitantes y turistas nacionales y extranjeros, pernoctaciones, duración de la estancia, y ***población flotante***. 

Para esto propósito, Kido Dynamics utiliza datos anonimizados de los operadores móviles, analizados dentro de un proceso que cumple de manera escrupulosa con el Reglamento General de Protección de Datos (RGPD) y que garantiza la absoluta privacidad de los usuarios. 

En este análisis, se considera que una persona pasa a ser residente de un municipio si ha pernoctado en el lugar al menos 15 noches de las últimas 30, con independencia de su lugar de residencia habitual para el resto del año. Esto permite identificar de manera adecuada los flujos poblaciones que se producen en distintos ámbitos del territorio nacional, y especialmente en ciertos municipios de interior y costeros, que pueden llegar a multiplicar hasta por 10 su población durante el período estival. 

![Poblacion flotante.jpg](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400450/articles_pictures/Poblacion_flotante_onwvb4.jpg)

Diana Andión, consciente de esta realidad de la población flotante, se puso en contacto con el equipo de Kido Dynamics con el fin de demostrar con datos que el índice de incidencia estaba sobreestimado y que el municipio corría el peligro de ser cerrado innecesariamente.

 El equipo de Kido Dynamics compartió los datos de la plataforma en los que se observaba efectivamente que el volumen de turistas durante el fin de semana del 3 y 4 de julio había alcanzado la cifra de 35’000, y que se había superado la cifra de 70’000 visitantes. 

![Conil 1.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400403/articles_pictures/Conil_1_k93xtu.png)

![Conil 2.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400404/articles_pictures/Conil_2_ewewj9.png)

El alcalde del municipio Juan Bermúdez se hizo eco de estos datos y los compartió con distintos medios en la rueda de prensa que concedió el 7 de julio a mediodía[10], justificando las razones de por qué el cierre perimetral del municipio era una medida demasiado severa para la realidad de la situación. 

Tras varias horas de reuniones con diversas autoridades donde se hizo hincapié en la necesidad de considerar estos datos a la hora de calcular la IA14, se anunció al comienzo de la tarde que la Junta de Andalucía decidía no cerrar perimetralmente Conil de la Frontera[11], dando como probado que efectivamente el volumen de turistas y visitantes constituyen un factor clave a la hora de calcular la IA14, especialmente en municipios como Conil con oscilaciones estacionales tan acusadas.  

En la rueda de prensa posterior, el alcalde Juan Bermúdez mostraba su satisfacción frente a esta decisión, y reiteraba la importancia creciente que el uso de los datos está adquiriendo tanto en la toma de decisiones como en la gestión de los recursos urbanos y municipales.

![Conil 3.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400406/articles_pictures/Conil_3_nsdkad.png)

El equipo de Kido Dynamics felicita a Juan, Diana y a todos los conileños y conileñas por esta ágil decisión y rápida actuación, que va a permitir al municipio y a los turistas que lo visiten disfrutar de una magnífica temporada de verano.

[1] [Confinamiento por la pandemia - Wikipedia](https://es.wikipedia.org/wiki/Confinamiento_por_la_pandemia_de_COVID-19_en_España)

[2] [link](https://www.nature.com/articles/s41586-021-03677-y)

[3] [link](https://es.wikipedia.org/wiki/Estados_de_alarma,_excepción_y_sitio_en_España#Segundo_estado_de_alarma_nacional_de_2020)

[4] [link](https://www.lasexta.com/noticias/nacional/preocupacion-aumento-incidencia-acumulada-coronavirus-cataluna-cantabria-registran-mayor-numero-contagios_2021070360e058d3dddcb20001b78729.html)

[5] [link](https://www.juntadeandalucia.es/boja/2021/541/5)

[6] [link](https://www.ine.es/dyngs/INEbase/es/operacion.htm?c=Estadistica_C&cid=1254736176951&menu=ultiDatos&idp=1254735572981)

[7] [link](https://www.ine.es/dyngs/INEbase/es/operacion.htm?c=Estadistica_C&cid=1254736177012&menu=ultiDatos&idp=1254734710990)

[8] [link](https://periodicoclm.publico.es/2020/05/15/farmacias-dispensan-marzo-abril-mas-350000-recetas-personas-tarjeta-sanitaria-madrid/)

[9] [link](https://www.cuencanews.es/noticia/72259/castilla---la-mancha/castilla-la-mancha-da-tarjeta-sanitaria-de-desplazados-a-300-madrilenos-al-dia.html)

[10] [link](https://www.europapress.es/andalucia/cadiz-00351/noticia-alcalde-conil-cadiz-pide-evitar-cierre-perimetral-seria-caotico-aboga-otras-medidas-20210707105251.html)

[11] [link](https://www.abc.es/sociedad/abci-conil-borde-cierre-perimetral-cuando-roza-plena-ocupacion-hotelera-202107071703_noticia.html)


Casos de uso del Big Data en el sector turístico durante el COVID-19: el ejemplo de Conil de la Frontera

I am happy to share here the results of our last Nature paper "Spread of a SARS-CoV-2 variant through Europe in the summer of 2020", an international collaboration with the University of Basel, the SeqCOVID-SPAIN consortium, and other amazing institutions in Switzerland, Spain and US.

Before the UK, Brazilian or Indian variants, there was a precedent last summer that put researchers as Dr Emma Hodcroft and her colleagues in alarm: a recently detected variant of SARS-CoV-2 (EU1) with a mutation in the spike protein was spreading faster than any other else along Europe. Early testing performed by Prof Iñaki Comas and Prof Fernando González Candelas helped to identify its origin in Spain in June 2020.

Fortunately, laboratory tests showed that the mutation in the spike did not grant a significant increase in the transmission or change in the antibody binding. So this result raised a question: how was this variant so successful all along Europe if there was no intrinsic difference in the virus?

Our data on international mobility between Spain and all European countries (the same we use for monitoring the tourism in any location in Spain) helped to build a model to correlate mobility with incidence estimates and created a map of how this variant emerged in Spain and spread to all the other countries: not all nationalities visit the same places or behave in the same way when they spend their holidays in Spain, and these differences are key to understand the incidence measured in the countries of residence. It also reflects the different policies taken by those countries with respect quarantines or testing, and how relaxed strategies opened the door for larger incidences.

![Countries.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400407/articles_pictures/Countries_rrhskw.png)

![COVID visitors.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400407/articles_pictures/COVID_visitors_xordvd.png)

In conclusion, we see how the intrinsic transmission property of the virus is not the only factor that makes a variant more successful over the others, but also our own behavior and travel patterns. Vaccination is a key differentiator for travelling in summer 2021, but we should keep tracking the evolution of the new variants and how they use our same airplanes and highways to move from one region to another!

This study also shows the relevant role of Spain regarding international tourism, being one of the major destinations in Europe and around the globe. Whatever happens in Spain will have its echo in the World! We really believe that data is the tool we need to understand the challenges that we will face in the post-COVID era.

Please follow [Emma’s thread](https://twitter.com/firefoxx66/status/1401833676317593600?ref_src=twsrc%5Etfw%7Ctwcamp%5Etweetembed%7Ctwterm%5E1401833676317593600%7Ctwgr%5E%7Ctwcon%5Es1_&ref_url=https%3A%2F%2Fmedia.licdn.com%2Fembeds%2Fmedia.html%3Fsrc%3Dhttps3A2F2Ftwitter.com2Ffirefoxx662Fstatus2F1401833676317593600url%3Dhttps3A2F2Ftwitter.com2Ffirefoxx662Fstatus2F1401833676317593600type%3Dtext2Fhtmlschema%3Dtwitter) for more details on this research and her amazing work!



The relevance of travel and behavior in the SARS-CoV-2 spread in summer 2020 and what we can learn for 2021

We announced a prediction one month ago about the forthcoming fourth wave in Spain. This prediction was based on the empirical observation that the number of COVID cases and deaths increase two and three weeks after the median of the radius of mobility crosses a critical value (70% of its pre-pandemic score).

The radius of mobility measures the dispersion in our displacements and captures the most relevant features in our daily mobility. We can see peaks in the radius for those special dates in 2019 (Easter, summer, national’s day, etc.) and valleys in the lockdowns of 2020 and 2021, as well as the peaks in the relaxation of the measures that systematically preceded each COVID-19 wave.

![Mobility radius.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400441/articles_pictures/Mobility_radius_grb7d8.png)

*Our team made an extraordinary effort collecting the daily radius of gyration of more than 10 million users, every day since 1st of January 2018 until yesterday, and updating it every 24 hours. We obtain a daily distribution from those users with very low to those with very high radius. We see here the evolution in time of the percentiles of such distribution, with the mean an the median. Thanks David Mateo, Jordi Bayer and Jordi Escrich for the effort.*

Indeed, we found that is correlated with the number of daily contacts based on the fact that most of our displacements are mean to meet with other people, or to do activities where there are also other people. Thus, more dispersion in our mobility is statistically correlated with having more contacts which is statistically correlated with increasing the number of positive COVID-19 cases. This chain of correlations makes the radius of mobility an earlier predictor of the evolution of the pandemic.

It is still early to confirm the accuracy of our prediction as data is still consolidating, but some comments in the media this week push us to make an update on this topic. This is our analysis of the present situation in Spain regarding mobility and COVID-19.

#### **1. Confirming the prediction from the radius of mobility**

The median of the radius of mobility crossed the critical value the last week of February, and the number of cases had the tipping point in the second week of March. Hospitalizations also had their tipping point three weeks later as expected (typically at the same week as for the cases or one week later). At the regional level, we see that the Community of Madrid crossed the critical value earlier, and thus all COVID metrics show an earlier increase in their values. For the rest of the provinces, cases and deaths were going down in almost all of the 50 main provinces in the week of our prediction. Last week, we had an increase in the number of positive cases in 46 of these provinces, and an increase in the number of hospitalizations, UCI cases and deaths in 20 of them. The situation is not good, and the radius is still above the critical value.

**Spain:**

![Radius Spain.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400454/articles_pictures/Radius_Spain_diw7ip.png)

**Madrid:**

![Radius Madrid.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400452/articles_pictures/Radius_Madrid_egw987.png)

**Legend:**

![Legend.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400426/articles_pictures/Legend_guyj3k.png)

There is a relevant data story behind these tables: The radius of mobility peaked during Easter and decreased the weeks after but still stays above the critical value. However:

1. Hospitalizations and ICU cases decrease in the last couple of weeks at the national level even if positive cases continue growing;
2. the number of deaths did not stop decreasing in all the period at the national level, and the tipping point in Madrid is 5 weeks after crossing the critical value. 
3. the growth in the number of cases looks slower than in the previous waves.

One candidate to explain this apparently contradictory behavior is the vaccination campaign. We expect positive effects indeed, but we expect them in the absolute values and not in the trends. We are facing here other effects related with the data itself, one related with the nature of the data and the other related with its harvesting process: scale invariance and retrospective consolidation.

#### **2. Interpretation of the data**

***Scale invariance***

When a system shows scale invariance, the natural way to check the data involves logarithmic scale. We learned with the pandemic the relevance of using the appropriate representation to see the data to get conclusions. Is the number of cases growing slower? Yes, in linear scale; no, in logarithmic scale:

![Scale invariance.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400462/articles_pictures/Scale_invariance_wq6wnf.png)

*The expansion of an epidemic is a multiplicative process and thus show scale invariance. Conclusions should be made by looking the numbers in logarithmic scale. When we do that, we find the growth of the third wave has the same rate as the previous one (right panel, dashed lines to guide the eye), what is not obvious in linear scale (left panel).*

In view of this plot, we can expect that the cases will skyrocket in the next weeks if it follows the same trend, as it looks like each peak and valley is more extreme that the previous one. Are we entering into a resonance?

***Retrospective consolidation***

The process of consolidating this data is complex and involves thousands of people. It is not easy, and it takes time, in some places more than others. Every day we get updates in the number of cases and deaths, but many of these cases are actually from days, weeks and months ago. It is also easier to consolidate a trend when numbers are very high since those complex situations have less relative weight to the total, but every contribution is relevant at low numbers as the weight of these fluctuations are higher in relative terms. In other words: it is harder to accurate update the data when numbers are low.

We can check how the time series change comparing historical data from different updates: confirmed cases are typically consolidated in a few days, but ICU and deaths may take weeks and months to get the actual picture. Madrid consolidates the number of deaths faster than the national level and already shows an increase in the last weeks (this is normal since the national level will update data as slower as the slowest region). 

![Scale invariance 2.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400459/articles_pictures/Scale_invariance_2_vfnpkb.png)

![Scale invariance 3.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400460/articles_pictures/Scale_invariance_3_nbvfkv.png)

*Number of cases are typically updated faster than number of deaths, that may take weeks or months to consolidate, as seen when we compare the updates at different previous dates.*

In summary, what we know now is:

- Regions with faster consolidation are already registering an increase in hospitalizations, ICU cases and deaths.
- Consolidation takes longer as the numbers are lower.

To understand better where we are now respecting the emergence of the fourth wave, let’s see what we knew during the emergence of the third one as data got retrospectively consolidated (let’s use logarithmic scale since we know it gets a better picture): 

![14 days average.png](https://shrouded-depths-78542.herokuapp.com/uploads/w_300/14_days_average_50e0c6ed52.png)

Spain crossed the critical value in its radius of mobility in the first week of December, peaking just before Christmas. The tipping point in the number of deaths between the second and third wave took place around mid December (according to consolidated data we have today). However, the 14-days average of the data we had available on the 28th of December just showed a constant decline trend. We only noticed a stabilization in the number of deaths some weeks later, around 15th of January, but the picture changed completely in only five days, when the effects of the wave in the number of deaths become obvious in the update on 20th of January. **This represents 7 weeks of difference between the early alarm of the radius of mobility and the consolidation of the data**.

#### **3. Conclusions**

*** Is this fourth wave growing slower?***

No, it is growing at the expected rate given the scale of the numbers, which are lower now than in the previous valley.

***Should we expect an increase in the number of deaths?***

We need to wait to consolidate the data, but yes, we can expect an increase. As we just shown for the previous wave, the picture dramatically changed in just 5 days with the new updates.

We should take decisions based on data, but we should care about where this data comes from and the bias it has. If we do not do that, we could spread misinformation. It is plausible that the vaccination campaign is now keeping the number of deaths low. And it is also plausible that the increase in mobility in Easter is not having any consequence in the number of deaths of COVID-19. However, it is highly probable that this is a superfluous effect of the data consolidation.

Fernando Simón (director at CCAES at the Ministry of Health) declared yesterday that “*Right now it seems that the effect that we could have expected due to the increased mobility in Easter is not occurring, according with the data that we have as of today*” [1]. He knows what he is saying, and he is right with “with the data that we have as of today”: the effect is not visible because the data is not consolidated.

Not consolidated data will always give the impression that the situation is much better that what it really is. We appreciate Fernando's effort in keeping the motivation high and recognizing the massive effort that the people is doing. However, this is sending a message of false optimism to the tired population that will push them to underestimate the danger of the fourth wave, and most important: burn and frustrate them psychologically once they realize that the fight is not over, over and over again.

This is known as the ***Stockdale paradox*** [2]: you should be optimistic in the long term to make it, but you will fail if you are naively optimistic in the short term, cumulating frustration after frustration.

Admiral Jim Stockdale was a war prisoner for 8 years at Vietnam and these were his words in his famous interview with author Jim Collins:

*Finally I asked, “Who didn’t make it out?”*

*“Oh, that’s easy,” he said. “The optimists.”*

*“The optimists? I don’t understand,” I said, now completely confused given what he’d said earlier.*

*“The optimists. Oh, they were the ones who said, ‘We’re going to be out by Christmas.’ And Christmas would come, and Christmas would go. Then they’d say, ‘We’re going to be out by Easter.’ And Easter would come, and Easter would go. And then Thanksgiving, and then it would be Christmas again. And they died of a broken heart. This is a very important lesson. You must never confuse faith that you will prevail in the end–-which you can never afford to lose–-with the discipline to confront the most brutal facts of your current reality, whatever they might be.”*

I am naturally optimistic (you do not start a PhD or create a start-up if you are not stupidly optimistic) and I know this situation will come to an end. But not yet, not now. We still need to hold some extra weeks.

Take care.

[1] [link](https://www.20minutos.es/noticia/4664409/0/tres-autonomias-mas-entran-en-riesgo-extremo-por-covid-pero-simon-es-optimista-hay-que-estar-satisfechos)

[2] [link](https://www.huffpost.com/entry/the-stockdale-paradox_b_5897ca82e4b02bbb1816bc38)


Update on mobility and COVID-19 data in Spain: The bias in retrospective consolidation and the Stockdale paradox

*Este artículo ha sido escrito en colaboración con el equipo de Kido Dynamics: Alberto Hernando, David Mateo, Bilal Khan, Juan Carlos Sánchez, Jordi Escrich, Raul Vallespín, Jordi Bayer, Luis Viotti, Carlos Roldán, Agustín Marrone, Pablo Lavallen, Antonio Andrade, Miquel Mateos.*

#### ***Análisis del exceso de movilidad interprovincial desde Madrid previo al Estado de Alarma***

Desde Kido Dynamics hemos realizado un seguimiento diario de los patrones de movilidad en España desde que se desató la pandemia provocada por el virus SARS-CoV2, o COVID-19. Como empresa de Big Data, creemos firmemente en que las decisiones deben estar basadas en datos y tratamos de hacer llegar ese mensaje a todos aquellos que desean tomar decisiones fundadas. Con este objetivo, hemos estado trabajado en continuo todo este año proporcionando métricas útiles para la gestión de la crisis en todos los sectores a los que ha impactado, tanto movilidad como turismo, comercio y por supuesto salud pública.

En un artículo anterior del 15 de mayo pasado [La responsabilidad es nuestra: correlación entre mortalidad del Covid-19 y movilidad](https://naukas.com/2020/05/15/la-responsabilidad-es-nuestra-correlacion-entre-mortalidad-del-covid-19-y-movilidad/), y en nuestro trabajo en colaboración con el grupo de J.J. Ramasco del CSIC, [Effects of mobility and multi-seeding on the propagation of the COVID-19 in Spain](https://www.medrxiv.org/content/10.1101/2020.05.09.20096339v2), mostramos cómo el número de residentes en Madrid visitando otras provincias relativo a la población de la propia provincia mostraba una alta correlación con la incidencia y mortalidad en esa provincia entre una y dos semanas después de analizadas dichas visitas. Por tanto, esta métrica (residentes visitando otras regiones) puede ser utilizada como predictor para el incremento y expansión de nuevos casos de COVID-19.

***Residentes en Madrid visitando otras provincias hasta la declaración del Estado de Alarma***

Dada la relevancia de la situación actual con la declaración ayer, 9 de octubre de 2020, de un nuevo Estado de Alarma, compartimos en este artículo resultados que consideramos de alto interés para su gestión. No es nuestro papel discutir sobre cómo se ha llegado a esta situación, pero sí lo es el contribuir con estos datos.

![Residentes Madrid.jpg](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/c_scale,w_765/v1644400462/articles_pictures/Residentes_Madrid_mo6mts.jpg)
 
En el gráfico superior mostramos el exceso de residentes en Madrid visitando otras provincias españolas desde el lunes 14 de septiembre hasta ayer 9 de octubre, día en el que se declara el Estado de Alarma. 

Lo que medimos con esta métrica es cuántos visitantes de más o de menos, cuya residencia típica sea Madrid, se encuentran en cada provincia durante las cuatro semanas anteriores, por tanto no se refiere exclusivamente a madrileños, sino a cualquiera que haya pernoctado una mayoría de semanas de forma continuada en la región de Madrid. 

La diferencia se calcula a partir de lo esperado para un día de la semana y un momento del año. Efectivamente, necesitamos corregir la tendencia estacional (fin de las vacaciones de verano y entrada en otoño) y el ciclo semanal (visitas de entresemana y de fin de semana). Es importante remarcar que no medimos viajes entre provincias, como se hace en el estudio del MITMA por ejemplo, sino visitantes presentes cada día en provincias distintas a la propia de residencia. Para el interesado, explicamos el proceso en detalle al final del artículo.

En una situación normal de movilidad, las fluctuaciones del exceso se mantienen bien acotadas (con pocas excepciones) en el margen entre -0.07 y +0.07, correspondiendo con subidas y bajadas de las visitas de un 7%. Es decir, si el valor esperado de residentes en Madrid visitando Alicante para un día entre semana de finales de septiembre son 24.000, el valor real observado será algo comprendido entre 22.320 y 25.680.

Sin embargo, hemos detectado en la última semana un claro incremento de estas visitas, indicando el movimiento de muchos residentes de Madrid a otras zonas de la península, creciendo hasta presentar su máximo el propio 9 de octubre. Algunas provincias como Ceuta y Melilla, o Valencia, Palencia o Segovia no muestran un comportamiento inusual, pero otras como Huelva, Cantabria, Asturias o Cádiz han crecido hasta cotas del 50%.

Para la totalidad de España, **el número de residentes en Madrid en otras provincias presenta un exceso del 11% para el 9 de octubre**. Dado que esta métrica ya fue identificada como un predictor en la evolución del COVID-19, consideramos de vital importancia tener en cuenta estos datos en la gestión del estado de alarma. 

Por supuesto, la situación hoy es muy diferente a como lo fue en marzo de este año, ya que ahora el virus ya está prácticamente extendido por toda la geografía española y se identifican varias regiones como puntos calientes, lo que significa que otros factores también serán influyentes en las próximas semanas.

***Residentes de Barcelona visitando otras provincias***

Hemos detectado también alta movilidad en el caso de los residentes en Barcelona. En particular, previo al estado de alarma, se detectan mayores oscilaciones y un exceso positivo en el fin de semana de la fiesta mayor de La Mercé del 24 de septiembre. También se detecta un exceso elevado en la primera semana de octubre con crecimiento hasta ayer 9 de octubre.

![Residentes Barcelona.jpg](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/c_scale,w_765/v1644400466/articles_pictures/Residentes_Barcelona_zgk5og.jpg)
 
***Anexo: Cálculo del exceso de visitantes***

Tomemos como ejemplo el caso de residentes en Madrid visitando Alicante:

![Residentes Alicante.jpg](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/c_scale,w_765/v1644400458/articles_pictures/Residentes_Alicante_wykynr.jpg)
 
Observamos para las últimas cuatro semanas una tendencia a la baja como resultado de la vuelta de vacaciones de verano, con cierta subida en los últimos días. Es visible a simple vista el ciclo semanal, con máximos durante el fin de semana y mínimos durante los días laborables.

En un primer paso corregimos la tendencia de fondo estacional, dejando evidente la señal del ciclo semanal. Se observa la subida en los últimos días, pero al coincidir con la subida esperada de fin de semana, necesitamos también separar este ciclo de la señal.

![Normalizado Alicante.jpg](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/c_scale,w_765/v1644400446/articles_pictures/Normalizado_Alicante_gxoduw.jpg)
 
Una vez eliminado el ciclo semanal es posible cuantificar el incremento del exceso, que en este caso llega hasta un 25%.

![Exceso Alicante.jpg](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/c_scale,w_765/v1644400420/articles_pictures/Exceso_Alicante_ydbsqk.jpg)
 
***About Kido Dynamics***

Kido Dynamics es una empresa de análisis de datos con sede en Lausanne (Suiza). La empresa opera en más de 10 países, y tiene como propósito la democratización del uso de los datos, la simplificación de las soluciones relativas a la movilidad, y la promoción de una economía que base sus decisiones en los datos.


Exceso del 11% en la movilidad interprovincial de Madrid el día de la declaración del Estado de Alarma

How much of the widely shared sentence “the virus doesn’t spread, we spread the virus” is true?

**Everything.**

According to WHO, the virus can spread up to a meter in an aerosol-generating procedure (another way to say sneeze) but generally falls just a few centimeters away from the host. However, it travels thousands of miles when we transport it. We are literally its mode of transportation, and jumping from one person to another is their multi-modal trip.

Other infectious diseases may have a physical location as the focus of infection, such as a water fountain with cholera, but we are the focus with coronavirus. Thus, the best way to limit the propagation is to limit ourselves from unnecessary exposure.

This is how many countries understood the situation, and the reason why they reacted with a lockdown. No one likes to have restriction of movement, especially in a time where we enjoy and use our freedom, but it was necessary.

**Is that true? Were the restrictions necessary?**

Yes. We have a clear evidence of correlation between mobility and mortality during the Covid-19 pandemic in Spain. This is not an opinion or a hypothesis, we measured it. In particular, we compared the mortality per province with the number of visitors per capita between each province and the focus of the epidemic (Madrid). We considered visitors in both directions: residents from Madrid visiting the provinces and residents from the provinces visiting the capital.

**What did we observe in the mobility data?**

If we compare the peak of mortality per province and the exchange of visitors per capita with Madrid one week before the declaration of the quarantine, we find a clear positive correlation:

![Visitors per capita.jpg](https://res.cloudinary.com/kido-dynamics/image/upload/w_320/v1644400469/articles_pictures/Visitors_per_capita_lycooz.jpg)

*Correlation between the peak in mortality due to Covid-19 with the number of visitors per capita exchanged between the province with the focus of the Spanish epidemy (Madrid). It is relevant to notice the high rate for the five neighboring provinces at the top right.*

A multivariate analysis tells us that 67% of the peak in mortality is explained by the mobility between the focus and the province normalized to the local population. It is not the only factor but it is the most relevant by far.

To put it in context: In the first week of March 2020, we have 6,000 residents from Soria visiting Madrid and 5,200 residents from Madrid visiting Soria. This represents 13% of the total population of the province. However, 37,000 residents from Barcelona in Madrid and 25,000 residents from Madrid in Barcelona for the same week represent just 1% of the total population of Barcelona. This brutal exchange of visitors relative to local population is what made Soria and Ciudad Real to have some of the highest mortality rates for Covid-19.

All these trips took place BEFORE the lockdown. These travellers were just doing what it was expected to do for that particular time of the year. For the five provinces bordering Madrid, we are talking about commuters who travel every day to their workplaces. These five provinces are indeed between the most affected by the epidemic, at the top right of the plot.

**Why Madrid?**

Finding the focus of the Spanish epidemic in Madrid is just a matter of probability. Indeed, Madrid is a European capital city with thousands of international travellers that spend days in the city or layover at the international airport. Using smartdestination.app, we observe around 100,000 international visitors each day just before the lockdown. This is not an unusual number and was the same for 2019 during the same dates. Madrid even had a peak of 200,000 international visitors during the 2019 UEFA Champions League Final. It is known that Balearic Islands host more than 300,000 visitors in summer, but at the beginning of March there were just 36,000. Only Barcelona compares with Madrid with 110,000 visitors for this time of the year.

It is interesting to find that the number of visitors in Madrid went down to 40,000 during the quarantine. This includes essential workers and those who decided (or had) to pass the quarantine in the city with the rest of inhabitants.

![Champions League.jpg](https://res.cloudinary.com/kido-dynamics/image/upload/w_320/v1644400398/articles_pictures/Champions_League_yxjlr5.jpg)

*Screenshoot of smartdestination.app for Madrid, where we can follow the evolution in the number of international visitors for 2019 and 2020. The fall after the lockdown is clearly visible.*

Once Madrid became the focus, the spread to the rest of the country just followed the expected patter when compared with national mobility. We can compare the pattern of trips between Madrid and the rest of the country with that of Barcelona and the differences are clear. With no measure to restrict in mobility put in place, the pattern of expansion was guaranteed to happen. 

![Thunders.jpg](https://res.cloudinary.com/kido-dynamics/image/upload/w_320/v1644400465/articles_pictures/Thunders_iitl8z.jpg)

*Number of trips in 2019 from Madrid (left) and from Barcelona (right) to the other Spanish municipalities. Geographical location strongly modulates the mobility pattern at the big scale.*

**Why did Covid-19 expand so fast?**

The most significant factor was the huge exchange of travellers between the capital city and the rest of the country. This triggered a multi-seed effect: there is no patient-zero as in classical models, but hundreds of unrelated focuses in a short interval of time. In the case of one patient-zero, the structure of the local social network may create bottlenecks that help to prevent the exponential expansion. Micro-targeting measures as those used by South Korea are useful to find each individual in that network and put them in quarantine. However, in a multi-seed scenario, there is no bottleneck and microtargeting is useless since there are already hundreds or thousands of potentially infected people. The only practical measures are general ones as restrictions in mobility and local lockdowns. 

![Camino a Soria.jpg](https://res.cloudinary.com/kido-dynamics/image/upload/w_320/v1644400396/articles_pictures/Camino_a_Soria_spahop.jpg)

Travellers exchanged with Madrid per capita (left) compared with mortality rates of Covid-19 (right). The pattern is clearly similar.

**How this result affects to the exit after lockdowns?**

Knowing the mobility patterns helps to identify risk areas without waiting one or two weeks for the rising of new cases. Correlation is not necessarily causality, but it does help improve our predictions. We can predict, within one week after the initial rise in cases at a location, which other locations may suffer a similar rise.

**Where did we get mobility data? Are we tracking people and contacts?**

We are not tracking contacts and this is important: we use aggregated data from mobile operators. The source of this data are not the smartphones but the telecommunication towers. We do not track GPS locations, we just count how many devices were connected to certain towers and how this number changes in time. Statistical methods help us infer the mobility with high precision in the totals, not in the individuals. We work with metadata: we know nothing about the content of the information shared in the network, just how many did use it. Thus, this is the most reliable technology available today in terms of privacy.

Working with this data is not easy. We talk about terabytes of data per day requiring scalable methods to clean noise and distil the relevant information for describing human mobility. After this treatment, we need to rescale number of devices to total population with statistical models. Even with the complexity of the methods involved, this data shows its predictive power and its potential to improve our cities, our travels, and our reaction to situations of global risk as this pandemic. All of it with no need of crossing the red line of privacy.

**What is the conclusion of this result?**

Restrictions cannot last for long or other important activities in our lives would be unattended. This means that streets and business will be opened again, as many countries are already doing. This also means that the responsibility is all on us now.

There are thousands of people who have no option but to commute or travel to other locations every day. We all have needs and responsibilities that push us to travel. No one can objectively judge if our reason for travelling is more important than those of others. We are all essential for someone else. However, we have no excuse anymore: we measured the effect of mobility on spreading the virus. Now we know it. It is not a hypothesis, it is an empirical evidence.

If the responsibility is on us, please do not abuse it. Let’s be responsible with our freedom of movement until we get the vaccine for the vulnerable.

**Reference:**
[**Effects of mobility and multi-seeding on the propagation of the COVID-19 in Spain**. Mattia Mazzoli, David Mateo, Alberto Hernando, Sandro Meloni, Jose Javier Ramasco (2020). ](https://www.medrxiv.org/content/10.1101/2020.05.09.20096339v1)


All responsibility is on us now: correlation between mortality of Covid-19 and mobility

*This article has been written by Kido Dynamics team, including Dr. Alberto Hernando, Dr. David Mateo, Jordi Escrich, Jordi Bayer, Raul Vallespin, Agustin Marrone, Luiz Viotti, Antonio Andrade, Carlos Roldán, Miquel Mateos, Bilal Khan and Juan Carlos Sanchez.*

**Almost two weeks after confinement was announced in Spain, citizens are responding very positively to these measures from a mobility point of view**

In our last article, we analysed how State of Alarm had affected mobility from a micro and macro perspective, bringing normal mobility values below 20% in some cases.

Here, we anticipate our conclusions:

Confinement measures (State of Alarm) announced by Spanish Government on March 13th have had a strong impact on people’s mobility patterns, with decreases ranging from 40% to 80%.

However, it looks like these measures are achieving or have achieved a plateau, where no further decrease in mobility could be expected coming days. It will be necessary to closely monitor how COVID-19 continuous to spread, since further measures might be required in order to limit even more people’s mobility, and hence decreasing likelihood of social contact, specially when looking at inter province figures.

After almost two weeks of confinement, we have analyzed if people are still responding to these measures and to what extent, from 6 different perspective:

1)    Micro-mobility within M-30 area

2)    Mobility within M-30 area

3)    Mobility between M-30 area and Madrid region

4)    Mobility between M-30 area and provinces

5)    Mobility within Madrid region (excluding M-30)

6)    Mobility between Madrid region (excluding M-30) and provinces

More details and daily updates can be found in https://macromobility.app or https://odmatrix.app, where we have included new features to improve customer experience. For further information, please contact us at  info@kidodynamics.com

***1)  Micro-mobility within M-30 area***

![Microtrips M30.jpg](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400439/articles_pictures/Microtrips_M30_ym7aj7.jpg)

As we saw in our previous analysis, micro-mobility has decreased an astonishing 62% if we compare March 9th with March 25th. This implies that 2 every 3 micro-trips are not happening within M-30 limits. 

Furthermore, we can see an additional 13% decrease between March 16th-18th and March 23th-25th, but how the trend is stabilizing, which might imply that confinement measures are achieving its maximum citizens’ response.

***2)  Mobility within M-30 area***

![Trips M30.jpg](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400468/articles_pictures/Trips_M30_gsniux.jpg)

A similar trend can be observed in those trips happening within M-30 limits, with an overall decrease of 80% and an additional 21% week to week decrease. 

Total number of trips maintains a continuous decreasing trend, with a day to day decrease of 6%, but certain hour ranges are stable, which might also indicate a maximum system effectiveness happening in the next few days.

***3)  Mobility between M-30 area and Madrid region***

![Trips m30 periphery.jpg](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400467/articles_pictures/Trips_m30_periphery_clr6er.jpg)

We can see again how communication between M-30 limits and rest of Madrid region has decreased up to 80% between March 9th and March 25th.

In this case, day to day decrease has almost achieved saturation, with just 4% additional decrease in traffic between March 24th and March 25th. Nevertheless, this percentage still represents 50’000 trips, and it is important to note that once a certain level has been achieved, further improvement become more complicated.

***4)  Mobility between M-30 area and provinces***

![Trips M30 provinces.jpg](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400468/articles_pictures/Trips_M30_provinces_ifwyn3.jpg)

From a confinement perspective, this metric and number 6 is incredibly important, as they determine the likelihood of COVID-19 keep on spreading across regions, since this is a direct measure of potential contact among people living in Madrid region and other regions.

We can see how these trips basically disappeared on March 21st and 22nd (weekend): we identified less than 5’000 people leaving this area, when a normal weekday register 90’000.

***5)  Mobility within Madrid region (excluding M-30)***

![Trips from:to Madrid.jpg](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400468/articles_pictures/Trips_from_to_Madrid_kzvk2v.jpg)

Madrid region (excluding M-30 area) holds on a normal weekday more than 10M trips. After confinement measures were announced on March 13th, we have seen a decrease of 44% in the number of trips happening, if we compare March 25th (4.6M) with March 9th (10.8M). Furthermore, it looks like this area is very close to saturation, with a day to day decrease of just 3% between March 25th and March 24th.

Considering the number of trips hosted in this area, and the limited reduction in terms of mobility, it might be necessary to consider further confinement measures depending on new COVID-19 cases in the region.

***6)  Mobility between Madrid region (excluding M-30) and provinces***

![Trips from-to Madrid 2.jpg](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400466/articles_pictures/Trips_from-to_Madrid_2_rm3jbw.jpg)

Finally, trips between Madrid region excluding M-30 and other provinces has experienced a 58% decrease compare to normal values. Considering that most of logistic and trading areas in Madrid, as well as many business areas are located in this region, this limited decrease indicates that an important economic activities and people’s exchange is happening between this region and rest of Spanish’ provinces.

Likewise, day to day decrease (24th to 25th) has saturated, this is, no further decrease in traffic has been detected between these two days, which might indicate that confinement measures have achieved its maximum effectiveness.

####*About Kido Dynamics*

Data is the new oil, but Big Data analytics remains a luxury good.

Kido Dynamics democratizes Big Data, providing the most advanced insights to make cities smarter. Companies, governments and public institutions can know deeply understand people’s mobility patterns: where they come from and go to, the routes they take, and time and seasonal aspects.

Our algorithms automatically and efficiently extract the most relevant information from telecom operator data, with a fully GDPR compliance technology.

Good quality, inclusive travel data are needed both to portray existing situations and to help identify problems related to the operation of today’s increasingly complex mobility ecosystem, and where current solutions just focus on a limited set of users such as micro mobility or ride hailing.




How much are people sticking to confinement measures?

*This article has been written by Kido Dynamics team, including Dr. Alberto Hernando, Dr. David Mateo, Jordi Escrich, Jordi Bayer, Raul Vallespin, Agustin Marrone, Luiz Viotti, Antonio Andrade, Carlos Roldán, Miquel Mateos, Bilal Khan and Juan Carlos Sanchez.*

***What will be the ‘new normal’ in mobility and tourism in the post-Coronavirus era?***

The First COVID-19 case in Spanish territory was detected on January 31st, in La Gomera (Canary Islands), and the numbers kept going south from that point on.

Fast forward March 9th, 1’231 cases were detected in the country. Regional authorities sprang into action kindergartens, schools and universities closed. Parents made quick adjustment to working from home and also to look after their kids. The dependable help of grandparents was not an option in this case because

a)     So far, the evidence suggests that Age has a direct correlation with the contagion and its severity, and 

b)    Kids are in general asymptomatic,

Companies and workers started a labor force experiment: working remotely, a common practice within certain sectors, but not broadly adopted in Spain, in general. This remote working would directly impact traffic in the main cities and productivity centers. 

Finally, on March 13th, Spanish Government declared the State of Alarm, that would apply on Sunday 15th 24:00  pm, which implies, among other measures, “*limitation of movement or permanence of people or vehicles at certain times and places or order them to meet certain requirements*”.

In loose terms, the State of Alarm implies to just keep operative activity and business of first need (supermarkets, pharmacies, hospitals, etc.) while closing down all other commercial activity. Mobility has a direct correlation with the spread of the virus and breaking it breaks the chain of contagion.

Although we are still far from assessing the impact of these measures at a socio-economic or health level, we at Kido Dynamics looked at the situation from the perspective of mobility data.

Here is what we see

***Normal Mobility in Normal Madrid***

With a population of 6.66 million inhabitants, Madrid is a big hub for both national and international visitors. More that 30M trips start or end inside its territory every single day and around 10M international visitors spent more than 20M nights in the city. Madrid is a vibrant city that never sleeps. 

**Commuters**

![Commuters.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400403/articles_pictures/Commuters_ykrdv9.png)

On a normal day, more than 3M commuters cross the city from home to work in the morning and go back home at the evening. Most of these people enter at work between 7-9 am generating a huge peak in the traffic. The return journey is staggered between 3 and 7 pm since many citizens go to third places for other activities in the evening. We can see this pattern during the week before the lockdown from Monday March 2nd to Thursday March 5th by filtering trips that started or ended in the province. This is the most expected daily traffic curve in most of the Spanish cities.

**Weekenders**

![Weekenders.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400478/articles_pictures/Weekenders_nev9uw.png)

This commuting pattern breaks at the weekends, starting on Friday 6th. Most business offer intensive morning schedules for their employees, allowing them to travel to their holiday destinations or family villages in the afternoon.

![Weekenders Mad Val.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400474/articles_pictures/Weekenders_Mad_Val_rkqegp.png)

This weekly migration peaks on Friday evening and Saturday morning, and peaks back on Sunday afternoon when they go back home to start the weekly cycle. This is clearly seen by filtering every trip starting in Madrid with destination Valencia at the East coast or Segovia at the near North.  (Please note that what we plot here are direct road trips.)

**Urbanites**

![Urbanites.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400474/articles_pictures/Urbanites_k1hkvg.png)

What happens inside the city in a normal day? We call it micro-mobility, i.e every trip that can potentially be covered with a mode of transport like a bicycle, electric scooter or just walking. These are short trips that typically cover basic needs or entertainment, or if you are lucky enough, your workplace. 

![Micro-mobility.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400439/articles_pictures/Micro-mobility_l9cjwh.png)

The micro-mobility pattern in a big city is modulated by what you find in that part of the city. For instance, when we analyse the area inside the physical barrier of M30 highway—with a mix of residential and commercial areas—micro-mobility has less structure with smoother peaks as compared with normal traffic and keeps almost constant for 12 hours from 8 am to 8 pm.

However, an area with hundreds of restaurants and pubs per square kilometer as that of the Downtown (Madrid Centro), micro-mobility leaves a special imprint. Indeed, the largest peaks are on the evenings of the weekend. Anyone who ever spend a weekend in Madrid knows. This is exactly what we see the weekend previous to the lockdown.

Detailed daily statistics on mobility and micro-mobility can be freely checked at https://macromobility.app at the province level and at https://odmatrix.app for any resolution.

**Visitors**

Madrid receives an average of 100K international visitors every day, staying an average of more than 2 nights, with a peak of 200K visitors during the Champions League Final last June 2019, as you can see in our daily statistics on tourism at https://smartdestination.app (free access). 

![Visitors.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400474/articles_pictures/Visitors_utgsgn.png)

Indeed, Spain is the second destination Worldwide in terms of number of tourists (just between France as 1st and US as 3rd) representing around 10% of country’s GDP, more than any other industry. Tourism is vital for any Spanish city, and any perturbation generates a deep impact in the economy and thus in people’s lives.

***Alarmed Mobility in a State-of-Alarm Madrid***

**The calm before the perfect storm**

The Spanish government monitored the situation in the week previous to the March 9th with no special measures, following the recommendations of the WHO. That changed after that day, and more and more business allowed teleworking and schools closed on Wednesday 11th. We see a smooth decrease in commuting traffic, day by day. 

![Daily commuters.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400413/articles_pictures/Daily_commuters_vxfru9.png)

Same effect on micro-mobility inside the city, where citizens reduced their exposure to the outdoors, driven by the in crescendo in the number of positive cases and deaths. Here is when most of the people understood that this was serious and their view of the crisis turned 180 degrees.

![Micro-mobility within Madrid.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400438/articles_pictures/Micro-mobility_within_Madrid_fkem3t.png)

**The sacrifice of the weekenders**

In view of the situation, the Spanish government declared the State of Alarm on Friday 13th. Thus, the first massive disruption affected directly to the weekend cycle. 

Some media shared the “news” that a tsunami of thousands of Madrilenians scape from the city heading to Valencia, Murcia or Alicante to spend some unexpected holidays and spreading the virus to other regions. Thanks to the data, we can tell this is Not the case. Just fake news.

**The impact on mobility was massive since citizens responded properly to the alarm, and this has to be properly acknowledged. This fast response saved many previous lives. Our deep respect to all these people.**

When we compare both weekends, the number of trips between Madrid and any other province reduced in a 71.5% and 73.3% on Saturday 14th and Sunday 15th with respect the previous week, as clearly seen when we compare the hourly traffic these days with Segovia and Valencia.

![Weekenders Mad Seg.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400475/articles_pictures/Weekenders_Mad_Seg_xcabir.png)

![Weekenders Mad Val.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400474/articles_pictures/Weekenders_Mad_Val_rkqegp.png)

We see these changes everywhere when we filter trips from Madrid to any other province. Taking into account that zero traffic is nearly impossible in order to keep the basic supply chain, these numbers tell that almost everyone respected the lockdown.

![Weekenders Interprov.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400475/articles_pictures/Weekenders_Interprov_qaihlb.png)

In addition, the intense night life of the Downtown just disappeared as restaurants were closed and people remained at home. There is no precedent in Madrid’s recent history for a situation like this. Let’s not forget these days are making History.

![Micro-mobility Mad Down.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400441/articles_pictures/Micro-mobility_Mad_Down_djo2xs.png)

**Back to work**

Monday 16th was the first day back to work for millions of people, while many others stayed at home with the kids since the schools are still closed. We see that the hourly curve remarkably changed as many schedules changed to intensive morning. People also reduced their exposure by commuting directly to home without spending time at third locations. 

![Daily Commuters Mad.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400413/articles_pictures/Daily_Commuters_Mad_ftzpkm.png)

Urban micro-mobility also suffered a major change, reduced in 50%. Citizens are indeed reducing their exposure to a minimum. But this is also a signal that a further virus spread might be happening, since there’s a clear movement of people among different regions.

![Micro-mobility within M-30.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400439/articles_pictures/Micro-mobility_within_M-30_bwctv6.png)

The effect is deep, and the crisis evolve to unprecedent states. **As public transport remains active, authorities need to fit this totally new and changing pattern with bus, metro and train schedules to reduce the exposure and travel time of the citizens.**

Looking at a larger picture, we can also see how interprovincials trips reduced more than 50%, which indicates that lockdown measures are being effective not only from a local point of view, but also from a regional point of view.

![Weekly Interprovince.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400478/articles_pictures/Weekly_Interprovince_y34c2l.png)

**What about the international visitors if borders are closed?**

The impact of COVID-19 in the economy will be estimated based on multiple elements, and tourism is definitely one of them. We cannot put right now numbers on the effect of flight cancellations and closing of borders in the economy, but we can put daily numbers on the international visitors at every Spanish region.

For Madrid, we observe in https://smartdestination.app how it start declining with respect to 2019 around Monday 9th, as also seen in other touristic cities like Barcelona. 

![Madrid tourists.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400429/articles_pictures/Madrid_tourists_h39def.png)

![Barcelona tourists.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400390/articles_pictures/Barcelona_tourists_yi0kwv.png)

Interesting enough, when segmented by visitor’s nationality, we detect that Chinese visitors were already missing since the Chinese government declared strong measures on travelling. This is also apparent for Italian visitors, with 30%/40% less visits as compared with 2019 for the same weeks, but we a slight increase last two days, probably motivated by Italian citizens approaching Madrid to fly back to Italy. Other countries also alerted to their citizens at the beginning of March, as we can see for Israel.

![Chinese tourists.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400398/articles_pictures/Chinese_tourists_yq7rkg.png)

![Italian tourists.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400425/articles_pictures/Italian_tourists_ytpel0.png)

![Israel tourists.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400425/articles_pictures/Israel_tourists_ki3hcs.png)

![AMerican tourists.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400386/articles_pictures/AMerican_tourists_nnk5hc.png)

All other nationalities declined along this week, as seen for instance with American and German tourists. 

![German tourists.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400423/articles_pictures/German_tourists_ge7y5g.png)

![South Korean tourists.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400464/articles_pictures/South_Korean_tourists_gzjzne.png)

![French tourists.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400421/articles_pictures/French_tourists_vge1f9.png)

Except for British. The number of British visitors in Balearic Islands continued as usual all that week as compared with 2019, only declining when their government recommended so once the lockdown was declared.

![Brits Balearic.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400396/articles_pictures/Brits_Balearic_oyruzn.png)

![German Balearic.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400422/articles_pictures/German_Balearic_bjipyt.png)

It is expected major extraordinary measures from the Spanish government next weeks, as the closure of hotels and airports. This. Is. Historical. Spanish tourism industry will suffer a total shootdown and reboot. 

***Conclusion***

What will be the ‘new normal’ in mobility and tourism in the post-Coronavirus era? Sorry for the clickbait, but we do not know and no one knows (beware of those who says they know the truth).

**We cannot predict the future and advance how the industry it is going to change in the post-Coronavirus era. However, we can provide daily updates of the state of things on mobility and tourism for fast response and educated plans. Let’s rebuild the new world with data-driven decisions.**

We are going to experience a ‘new normal’, where all our previous knowledge or hypotheses will be obsolete. All trends and forecast just burst. 

In this context, it will be critical to have real time information about how things are evolving and react fast for choosing the right decisions. The situation changes every day turning long term strategies into short term guerrilla, and any information will be incredibly valuable to navigate these uncharted waters.

Data, massive, scalable, reliable, and available, will be a competitive advantage for those embracing the new normal.

#### About Kido Dynamics

Data is the new oil, but Big Data analytics remains a luxury good.

Kido Dynamics democratizes Big Data, providing the most advanced insights to make cities smarter. Companies, governments and public institutions can know deeply understand people’s mobility patterns: where they come from and go to, the routes they take, and time and seasonal aspects.

Our algorithms automatically and efficiently extract the most relevant information from telecom operator data, with a fully GDPR compliance technology.

Good quality, inclusive travel data are needed both to portray existing situations and to help identify problems related to the operation of today’s increasingly complex mobility ecosystem, and where current solutions just focus on a limited set of users such as micro mobility or ride hailing.




Towards a new normal in mobility?

(Preamble Note: most of the content of this article is either common knowledge, or strong assumptions broadly accepted by academics and professionals. What is totally unique is the way we are actually somehow confirming that knowledge).

Spain has been for centuries a crossroad of cultures and people, starting with Phoenician, Greeks, Romans, Germanic tribes, Arabs, etc. Without diminishing their impact, in Spain, as well as in most of European countries, Romans left the most relevant and permanent footprint. We have heritage not only our language, but our law, sewage and hydraulic systems, harbors… and one element which still defines the way we communicate and interact with our national peers: road networks.

![Mapa romano.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400433/articles_pictures/Mapa_romano_xvbxbu.png)

Figure 1. Mains roads in Hispania. Source: WikiMedia Commons. Author: RedTony

When Roman Empire conquered a new land, they would rush to build a meaningful and dense network of roads for different purposes, but mostly for (not exhaustive):

Provide the army with a quick access to any point of the territory in case of riots and/or war
Improve the flux of communication and information, including Latin knowledge spread
Allow a more efficient commercial activity and assets exchange among regions
We can undoubtedly recognize that Romans had a clear view about how infrastructure would play a strong integrative and cohesive role in their territory.

***TWO THOUSAND YEARS AFTER***

Living in the geek economy, in a multiconnected society, with low cost flights, 5G connectivity, 4K streaming… we might think about Roman economic model as old fashion. Or maybe not.

At Kido Dynamics, we are driven by a strong academic and research purpose, not only to make our solutions meaningful, relevant and useful for today’s society, but also to understand how certain decisions that we might be taking today could impact and define the future of our children tomorrow… to the same extent that we are still impacted by what Romans did 2’000 years ago.

In that context, we have completed the ***largest and most extensive mobility analysis ever done Worldwide*** in a given region, in this case Spain, with more than ***4 billion trips*** analyzed over the last 12 months [1]. Most of the breakthroughs we have unveiled will be publicly available in the coming weeks, but I wanted to share some interesting insights we have discovered while diving through terabytes of data.

***THE SOCIAL NETWORK PERSPECTIVE***

In today’s digital ecosystem, social networks are an essential component of our day to day, and we concentrate a big share of our social interactions in the digital sphere. 

Furthermore, at Kido we develop most of our day to day work on a pure digital context (actually we don’t even know personally most of our clients). But we are physicists and engineers by trainee, and we generally need evidence about how accurate our assumptions and models are versus reality. (Just in case you missed what we actually do, we describe and predict human behaviour applying mathematical methods and procedures inspired in physics, as the concept of maximum entropy in thermodynamics or the eigenstates and propagators of quantum mechanics [2]).

So, what if (certain) human behaviours could also be explained with other models, such as ***social networks*** or virality effects?

In this context, we defined a theorical framework that would allow us to simplify some of our initial assumptions and extract relevant insights from our massive set of data. Since P2P communications would be too granular for a broad base analysis (people interact in average with a maximum of 150 people [3]), we decided to aggregate by administrative units: ***municipalities***. Sample size: more than 15M users distributed in 8’131 municipalities in Spain.

As described previously, the base of our work was the number of trips that people from every municipality performed over the 12 months of 2018, so we classified all municipalities based on the number of trips that each municipality generated or received from any other municipality. Under this context, we play the next game: we define that a ***municipality*** is a ***follower*** of another as long as there were more than 10 trips originating in its territory with the second municipality as destination. The more municipalities connected, more followers, and more popular. In this social network, who are the influencers[4]?

![Followers.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400422/articles_pictures/Followers_uu0thz.png)

To our surprise, more population does not necessarily implies being popular. There is a correlation indeed, but we can find remarkable cases such as San Fernando de Henares which is higher in the ranking than other capital cities with more than six times its population. A reader familiar with the Spanish geography will immediate find that San Fernando, as a satellite city of the country’s capital (Madrid) has a privileged location both geographical and in terms of infrastructure. And this is totally true. However, this same reader would also think that L’Hospitalet (55th, 2’532 followers), a member of Barcelona’s metropolitan area with a density of population double than that of Manhattan, should also be in the top of influencers. But it is not. The underlying dynamics is complex and sometimes unexpected.

If we represent the total number of trips received per municipality, we find interesting patterns:

![Spain network.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400468/articles_pictures/Spain_network_ogvx3z.png)

Figure 2. Heat map of the number of trips for every municipality. Source: Kido Dynamics

- Municipalities with highest number of followers are located along main roads or close to infrastructure hubs.
- Full regions in Spain condensate the lower rankings in the number of followers.

It is well-known that the area of municipalities in the south are bigger than those in the north for historical reasons (same effect as the US counties from East to West) so one may think that it is related with the fact that the bigger the area, the bigger the population. But things, as usual, are more complex: this is the so call España vacía [5]or Empty Spain, as can be seen in the population density map of the country below. Ironically, Spain is one of the European countries with the lowest overall density of population, but it also owns the city with the largest density of population of all Europe. Spain is complex.

![Spain Population.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400469/articles_pictures/Spain_Population_hkm6jv.png)

Figure 3. Spanish population density map. The empty Spain vs the peaks of population are clearly visible. Source: Kido Dynamics

Since we had a nice level of disaggregation, we decided to dig deeper and analyzed for multiple municipalities of difference size, not only the number of connection or followers they had, but also how these followers were distributed across the geography.

***Case 1: Madrid***

![Madrid network.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400432/articles_pictures/Madrid_network_eisdcx.png)

Figure 4. Connections between Madrid and the rest of the territory in Spain. Source: Kido Dynamics

Not many comments in this case. The largest metropolitan area in Spain is systematically connected with any point of the territory, which is also favored thanks to its central position.

***Case 2: Barcelona***

![Barcelona network.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400393/articles_pictures/Barcelona_network_j2mc4a.png)

Figure 5. Connections between Barcelona and the rest of the territory in Spain. Source: Kido Dynamics

The second largest metropolitan area in Spain, has strong bias in terms of communications towards surrounding heavily populated regions such as Valencia, Alicante, Zaragoza and Madrid, while exchange and trips the West part of the country is barely testimonial.

***Case 3: Bilbao***

![Bilbao network.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400389/articles_pictures/Bilbao_network_i1hkr0.png)

Figure 6. Connections between Bilbao and the rest of the territory in Spain. Source: Kido Dynamics

An important industrial pole in the North with strong influence in the surrounding region (Basque country, Navarra, Cantabria, Palencia, Burgos) and strong communication links with Madrid, Zaragoza and Valladolid.

***Case 4: Orense***

![Orense network.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400448/articles_pictures/Orense_network_wllyy1.png)

Figure 7. Connections between Orense and the rest of the territory in Spain. Source: Kido Dynamics

Orense represents a typical regional hub, with strong connections with surrounding region, but very limited (or inexistent connections) with the rest of the territory.

With a difficult orography, Galicia region has always been somehow isolated from the rest of Spain, developing its own local economy ecosystem. It’s important to note that many of these out of the region connections are happening across and along three main communication axis: AP-6 (Valladolid and Madrid), A-8 (Asturias) and A-66 (Gijón/Badajoz).

 Case 5: Siruela

![Siruela network.png](https://res.cloudinary.com/kido-dynamics/image/upload/w_300/v1644400464/articles_pictures/Siruela_network_kod4hx.png)

Figure 8. Connections between Siruela and the rest of the territory in Spain. Source: Kido Dynamics

We chose for out last example a 2’000 inhabitants’ village in the so call Empty Spain. We can clearly observe that:

1. Number of communications decrease enormously, with no municipalities receiving  over 1’000 accumulated trips.
2. Most of the trips are concentrated in few municipalities surrounding the village

***Important to note:*** travel distance from Siruela to the closest speedway (nearby Don Benito) is 90 km or 60 min, while travel distance to Madrid is 3 h 20 min for 274 km. Similar time to Madrid-Valencia with almost 100 km more.

Not surprisingly, Siruela’s neighbors visit Madrid, Badajoz (2h 12 min) or Córdoba (2h 13 min), but not even consider to visit Sevilla, Valencia, Barcelona or Bilbao. 

***FINAL CONSIDERATIONS***

It is important to appreciate how infrastructure acts as a territory integrator, facilitating access to multiple locations in a safe and fast way. There are multiple studies and analysis regarding this phenomenon, but in most of the cases are based in a limited/reduced sample and concentrate in a small region or metropolitan area.

To our understanding, this is the largest and most exhaustive analysis ever done in a country to understand how infrastructure layout underpins communications, trade and exchange among municipalities, regions and ultimately, people. (Utilising number of trips as a basic, simplify proxy for this analysis)

In that context, we are still bearing what Romans decided 2’000 years ago, since most of existing speedways are just reflecting the ancient Roman calzadas, with better bridges, tunnels and security, but they still connect the same points defined 20 centuries ago.

We need to ask ourselves if many of the decisions related to infrastructure and mobility investment we are taking today are based in what we consider optimal for today’s needs, or if they are still based in assumptions and mindsets that we have inherited, while we have a massive, powerful and limitless set of data which might help us to optimize (and rethink) this process.

**So what will you do next time you have to invest billions in a new speedway to dynamize a region? Trust Romans or trust Data? ** 

[1]At this point, I will omit any explanations related to anonymization, aggregation, privacy and GDPR compliance for the sake of simplicity. If you want to know more, please post me. I will just say that our analysis is based on more than 15M users, and that peer to peer communications are not considered or even analyzed, since they are not relevant/useful for our purposes

[2]The workings of the maximum entropy principle in collective human behavior. A Hernando, R Hernando, A Plastino, AR Plastino. Journal of The Royal Society Interface 10 (78), 20120758

[3]Unravelling the size distribution of social groups with information theory in complex networks. A Hernando, D Villuendas, C Vesperinas, M Abad, A Plastino. The European Physical Journal B 76 (1), 87-97

[4]We do not include those municipalities with less than 300 inhabitants (which are mostly those with less than 10 trips per connection) to protect the privacy of the users by working with large aggregated values only.

[5][link](https://www.geografiainfinita.com/2019/03/un-recorrido-por-la-espana-olvidada-y-vaciada/)

How would Spain look like as a Social Network?

I attended last week in Stockholm (Sweden) the UITP Global Public Transportation Summit. This represents a major event for Worldwide transportation players, and I enjoyed talking to experts from all over the World and get different hints about what's going on in this industry. I was not specially interested in hardware providers (even in Mercedes Benz new coaches are stunning) or ticketing solutions, but mostly on software and solution providers.

####  **What happened during UITP Summit in Stockholm last week?**

Most of people in Sweden were talking about Mobility as a Service (**MaaS**), this is "*a shift away from personally-owned modes of transportation and towards mobility solutions that are consumed as a service. This is enabled by combining transportation services from public and private transportation providers through a unified gateway that creates and manages the trip, which users can pay for with a single account. Users can pay per trip or a monthly fee for a limited distance. The key concept behind MaaS is to offer travelers mobility solutions based on their travel needs.*"

I immediately thought about some fresh news released recently about dockless scooters and the **impact** of these theoretically environmentally friendly solutions (some companies were marketing dockless scooters at the venue). Actually, dockless scooter releases more CO2 than a mid-size private car, while they have a life span of merely **30 days**. But also all investment hype around **e-scooters** and e-bikes, while the actual impact in mobility is at least **unclear**. Bear also in mind my previous newsletter regarding how many additional miles Uber or Lyft were adding for every private mile they remove.

I rapidly connected the dots in my mind, and I remembered what happened in China just a couple of years ago, and the environmental impact that this phenomenon generated.

***So this is my conclusion:*** none of these players is actually providing a sustainable solution for mobility, basically because they don't have a clue about mobility. They are just taking advantage of an investment hype, presenting mind blowing numbers with very little substance, and expecting either to become the monopolistic player, or to be acquired for a larger firm, or to keep on burning OPM (other people's money), waiting for a future breakthrough, name it autonomous vehicles, flying cars or whatever, while running an unrealistic and never profitable business model. Have you heard about UBER??  

#### **Any other interesting insights from UITP Summit in Stockholm?**

It's interesting to see how business intelligence (BI) has taken over almost every sector and industry Worldwide. While I was in Stockholm, I read about SalesForce acquisition of Tableau. My opinion about BI tools is strongly biased, since they generally used well structured data to run and obtain powerful insights, and we systematically play in a World of unstructured data. Nevertheless, I can see the value of such an acquisition.

Interestingly, many companies including big fish such as Alstom or Siemens, are developing their own BI tools for the data their systems are continuously generating, from network performance to train occupancy to trip time to ambiance temperature to... everything in a fancy, colourful dashboard. Again, I appreciate having all this information in a single place, but I also wonder why they did not have developed such a tool before. Many start-ups are trying to tap into this opportunity, providing tailor made solution to mobility operators. There is in general a lack of differentiation among solutions, and in most of the cases users could shift without incurring in big costs. Just my opinion.

#### **What about competitors at UITP Summit in Stockholm?**

There was one company which got my attention for different reasons. We shared a panel about big data applied to infrastructure and they exposed their solution and how they deploy it. In a nutshell, they have developed a SDK that they install in third party apps. This SDK gathers all mobility data generated with your mobile, while they can identify how you are moving around and where you are.

Those applications with this SDK installed (few thousand of users), will require users' permission to share location and speed with the company, I assume that everything within a GDPR framework. But as we saw in the presentation (nothing new by the way) this information can be incredibly invasive, identifying for example a what time and in which place you are picking your kids from school.

Company expressed that this information leaves the devices anonymised and aggregated (???), but they did not clarify what happens if someone is eavesdropping. 

Of course, the level of detail they are able to obtain is amazing: street level, per second accuracy, multi modal transportation, etc., but I wonder: at what price?

***My conclusion here:*** out there, there are comparable solutions to ours, but... I have not seen any solution that can compete with ours matching privacy, accuracy, scalability and global reach.



Kido Dynamics - Blog