INNovATIoN WITHouT BoRDERS
Engineering school recognized by the French Governement since 1922
Diploma accredited by the French Commission on Engineering Credentials (Commission des Titres d’Ingénieur - CTI)
Member of the Consortium of the Top French Schools in Engineering and Business (Conférence des Grandes Ecoles - CGE)
Member of the Union of Independent Grandes Écoles (UGEI)
ESME Sudria alumnus, Raphael Presberg (class of 2018) is the co-founder of start-up Eiffo, incubated within IONIS 361. With his partner Alexis Tuil, he participated in the launch of Data Against Covid-19, an open source initiative that aims to better inform the evolution of the epidemic by region in France and which now has several hundred volunteers.
With Eiffo, Raphael Presberg and Alexis Tuil developed a predictive “customer health” score for Saas B2B companies that sell products with annual or monthly subscriptions. “This score allows the customer success teams of these companies to obtain objective visibility of their client portfolio through the data and to identify in advance the clients at risk in order to adapt their strategy”, explains the data and artificial intelligence engineer who embarked on the entrepreneurial adventure in January 2019. For all that, it is today for another adventure, citizen this time, that the ESME Sudria wished to speak with this Alumnus: it is Data Against Covid-19.
How did this idea of Data Against Covid-19 arise ?
Raphael Presberg: Everything started from a publication posted on Linkedin by Lior Perez, the Software Development & AI Lab manager at Meteo-France: we were then at the beginning of March and he called on Data Scientists to collect data. With Alexis, we decided to join him and, initially, there were only four of us. Our goal was then to try to start collecting data on the spread of Covid-19 in France by region, which did not exist. To do that, we would go through the regional and provincial newspapers… In short, all local news reports on the subject, to get information on the affected persons. The objective was to build a fairly granular data source, with a lot of information about the patients affected, so that we could then do some analysis on it.
When did the project start?
About two weeks ago. When we started, I took over the Grand Est region, where there were between 100 and 120 cases, and Alexis was responsible for the Île de France region, which was not very affected. Today, these figures have exploded. But counting the number of cases does not mean anything anymore because, in France, people are not tested in their majority. Without testing, the number of confirmed cases can always be recorded, but the total number of actual cases cannot be counted.
So has the initiative evolved rapidly?
Yes, along with the number of people involved. Very soon after the launch, other volunteers joined us. From four, we went to eight. We then created a small discussion group via the Slack app. At the same time, the number of patients increased at very high speed and it was therefore no longer possible to obtain information on these cases in a granular manner. Instead, we chose to collect data to not describe in detail the profiles of those affected, but to track the number of people affected by the disease by region and day. We trained together and pooled our work to achieve something solid. By March 8, there were twelve of us on the Slack to consolidate this data. And then, without knowing how, two or three days later, we found almost 450 “data guys” on the Slack, all ready to help! This immediately led to the creation of two data dashboards which enabled us to follow the evolution of the Covid-19. Today, the group continues to grow and we have to be over 900 volunteers. This leads to the development of many projects.
In addition to mapping the pandemic, what kind of projects are carried out?
We soon came into contact with state institutions, for example through people from Etalab, the organization whose primary mission is to digitize public institutions. We also approached the Directorate-General for Health (DGS) to understand what data it collected and sent to Santé publique France (French Public Health). Then there was some lobbying to make that data public so that we could work on it and compare it with the data we had put together. Alexis then spoke with epidemiologists and doctors to understand the needs. All this made it possible to launch new projects, mainly prediction projects. This concerns the prediction of the number of ICU beds in regions, the prediction of the epidemiological peak per department, the prediction of the number of masks, gels and people needed, the prediction of the number of real and cases tested, etc.
So, does this citizens’ initiative helps the authorities?
Yes. As a matter of fact, one of our dashboards is on the government website today. This was obviously made easier by the presence of Etalab people at our side who justified our credibility.
The aim is to help, of course, but do you think, this could be also a good way to make the general public understand the importance of data’s role, in respons to major issues at a national level?
It is above all a means of providing information in a purely quantitative way. The creation of these dashboards is also designed for the general public, or rather the general public already familiar with curve analysis, to enable them to understand the evolution of the epidemic.
Is your work also based on comparative analyses with past epidemics?
Initially, this was part of the projects we had, especially with the figures on the spread of the Ebola virus that Alexis had been working on. But in the end we were not able to implement it because we don’t have all the information and hindsight we need today. In order to analyse an epidemic, you need sufficient hindsight and data.
Do you have a website that allows to track your work?
We’re working on it right now. With the strong growth of our Slack Task Force, it seemed natural to us to create an internal and external communication team. We will soon be able to be present to communicate our work on social networks. In the meantime, we still have a website with the address of our Slack and our contacts (https://bzg.fr/covid19-developpeurs-datascientistes-comment-aider/).
So if developers, data scientists or community managers want to join the movement, they can still do so?
Of course! These profiles are welcome. We also need Project Managers, lawyers, data protection specialists… This is a truly open source initiative and everyone can come and help, with their skills, ideas and desires.
Finally, have you received direct requests from organizations to work on certain solutions? Can they also contact you?
Not yet. However, the AP-HP and Santé publique France have asked if, in our working group, people on fixed-term contracts or freelance could be available for missions to be carried out. On the other hand, we are forwarding our results to the DGS to challenge their models and its evolutions.
Share this News:
About ESME Sudria
Founded in 1905, ESME Sudria trains multidisciplinary engineers, important professionals in the sectors of the technologies of the future: energy, systems and environment; on-board systems and electronics; images, signals and networks, digital intelligence and data. Its modern teaching method is derived from three components: the importance of innovative spirit; the omnipresence of projects and initiative; and a resolutely international, human, and cultural outlook. Since its creation, more than 14,000 engineers have graduated from its courses. ESME Sudria delivers a Master of Engineering certified by the Engineering Education Commission (CTI) and recognized internationally as a Master of Science.