CovidTracker, the success of a data savvy

CovidTracker is a website visited by tens of thousands of people every day. It allows you to visualize all the available data related to the covid-19 pandemic in France in the form of graphs updated every day. An unprecedented success initiated in Nancy.

At the beginning of CovidTacker, there is Guillaume Rozier. A young man with a passion for data. His thing is all this information available, which we do nothing and which should nevertheless be used “before making a decision“.
It is of course on his personal time, in full first confinement, that he developed this digital tool. Initially, he only wanted to share it with a few friends on social media. But very quickly, its graphics were spotted and used by editorial staff, but also by hospitals. Encrypted data is quite stodgy to use.
After a few months, his twitter account explodes. With the latest tools, which he recently added to his site, he has crossed the threshold of 15,000 subscribers. In recent weeks, its pages have garnered millions of views. The site is ad-free and completely free.

I never imagined having five or 6,000 followers on Twitter. At the end of spring, I said my goodbyes. And now I have 15,000 subscribers.

Guillaume Rozier founder of CovidTracker

A time-lapse that allows you to see the evolution of the situation over 3 months

A time-lapse that allows you to see the evolution of the situation over 3 months

© CovidTracker

To see this time-lapse (Ultra-accelerated effect produced frame by frame).

Tools, statistics and digital technology that arouse enthusiasm

In recent weeks, Guillaume Rozier has developed in addition to dashboards for each department and for each region, four new tools using his algorithms.
CoviDep is a tool that presents the departments that have exceeded the number of hospitalizations for Covid-19 from the peak of the first wave and those that have not yet exceeded it.
The wave comparator makes it possible to compare the height of the two different epidemic waves of Covid-19: that of spring 2020, and that of autumn 2020. Three variables are compared: hospitalizations, resuscitations and hospital deaths.
The Deconfinement Calculator allows you to calculate the duration of the confinement according to the value of the Coronavirus reproduction rate over each week. The higher the reproduction rate, the longer the confinement will have to last.
The risk calculator is the most successful one: “it makes it possible to estimate the risk that a person is positive in an event. For this, we need two pieces of information: the incidence rate which is the number of cases over a week per 100,000 inhabitants. And the second piece of information is the number of people attending the event. The incidence rate is automatically filled in from the moment you indicate your department … “Guillaume Rozier insists:”It is not the risk of catching the virus that is indicated, but the risk of being in the presence of a positive person… You may be in the presence of a positive person, but if everyone is carrying a mask and that the other instructions are followed, there is no more risk“.

During the epidemic peak, there were 100,000 people connected simultaneously to the site. In normal times, these are several tens of thousands of users per day.

Guillaume Rozier

Guillaume Rozier, a data enthusiast above all

In normal times, these are several tens of thousands of users per day.
For Guillaume Rozier, who continues to work there several hours a week in addition to his work, it is above all a passion. “I am interested in the whole data chain, from the collection of digital data to visualization at the end of the chain, including processing, and even storage. The treatments can be simple or go through complex treatments such as artificial intelligence, machine learning. This is not the case with CovidTracker“.

Today in his professional life Guillaume Rozier is a “data scientist” in Paris.
It’s in an engineering school in Lorraine, Telecom Nancy that he took a liking to data: “Telecom Nancy allowed me to have the skills to set up its tools. The whole data processing part: how to make algorithms to download data, to transform them, to generate graphs … From 2016 to 2020. Normally, it’s three years. Me, I stayed there for four years because I had a double degree“.
He specialized in Big Data applied to the biomedical field.

CovidTraker charts with data from Public Health France
CovidTraker charts with data from Public Health France

CovidTraker charts with data from Public Health France

© CovidTracker

I don’t see how you can make a decision without data. Data is a gold mine. Raw data is unnecessary.

Guillaume Rozier

Using data is therefore obvious for Guillaume Rozier. But he insists: not just any old way “by respecting all the rules of data anonymization and the protection of privacy. It is completely possible within the framework“.
They must also be put into perspective in their context. He willingly tells the story of a journalist publishing raw data taken out of context. “He published the raw figure of the number of deaths on a Monday in full deconfinement. This number is always higher on Monday for different reasons, it also contains the numbers for Sunday etc. While we were in a period when it was only going down. What he wrote was true in gross value, but false, because the data is not put in its context. If he had made a graph, in two seconds he would have seen the reality of things“.

The data is all verifiable

Guillaume Rozier “only” recovers existing data. Everyday, Public Health France publishes figures related to the pandemic. INSEE and other sources give it the number of beds available and the number of beds occupied. Thanks to the algorithms he has developed, all of these numbers this daily spill into his computer which digests them and makes them visible as graphs. He has the idea of ​​crossing some of these data, for example the occupied beds and the available beds and by this simple reconciliation of the figures, he highlights the essential information to understand the different situations. He even sometimes reported column reversals to his source on his post, because the graph that appeared was totally improbable.

I make choices when I make a graph. I make choices of color, shape, size.

Guillaume Rozier

Occupation of resuscitation services in the Grand Est
Occupation of resuscitation services in the Grand Est

Occupation of resuscitation services in the Grand Est

© CovidTracker

I make choices when I make a graph. I make choices of color, shape, size. The simple act of changing the scale can influence the perception of things“.
In saying these words, Guillaume Rozier is aware that his tools are now used by many professionals. With its graphics and time-lapses, it makes this pandemic visible. He knows they are just tools and nothing else. He receives many messages of support and encouragement and that is enough for his happiness.

Share on facebook
Facebook
Share on pinterest
Pinterest
Share on twitter
Twitter
Share on linkedin
LinkedIn
Share on email
Email

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.