3 Questions To… Alexis Thual, an open source developer and machine-learning engineer. Previously founder at Arkhn, an organisation which aims at building open-source data integration systems for hospitals, and developer at Bureau Ouvert's at the French National Assembly. Alexis is among this new generation of European makers that strives to process Data for Good. During this time of crisis, he just developed with Raphaëlle Bertrand-Lalo a platform https://coronastatus.fr to help having a better idea of how many real Covid-19 cases are currently present in France - and in the world - and provide healthcare facilities with insightful data and predictions. Coup Data wanted to know more about it and ask him his view on how open source technology can contribute to enhance public services.
1.
You just developed a solution called "Corona Status" for confined population to produce anonymised and open data about their current medical status. Can you tell us more about it?
For now, the only goal of this application is to help build the dataset you mention. The latter is meant to help tackle two main issues. First, our estimation of the current number of real COVID-19 cases is mainly based on the number of deaths amongst diagnosed people ; it is therefore highly biased to our testing policy, and thus rather imprecise. Second, once our healthcare system overloads, it is very hard to keep track of how the pandemic spreads.
If enough people use Corona Status, one should be able to better estimate the current status of the pandemic, without running more medical tests, and even if our healthcare systems runs critically overloaded. This dataset is meant to be open, which I reckon will allow a lot more people to produce interesting models and analysis about this topic. However, for this dataset to be considered anonymised, I will need a lot of individual answers.
2.
How do you believe open source technology can help building better public services?
I believe open source technology is, by essence, meant to work well with public services, since they allow for great transparency and contributions from a great variety of people. From what I have seen so far, public services feel safer relying on proprietary software for key features (architecture, data storing) which makes sense from their perspective, even though I believe open source architectures and softwares (with Ubuntu, Mozilla) have proven to be production ready.
« Furthermore, I would claim that open / standard data is even more important in 2020 than open source is. »
On top of this, I think a lot of fine-tuned projects - anything related to the parliament, climate change, etc - would gain a lot of credit being open and could sometimes be maintained by publicly-funded teams. Furthermore, I would claim that open / standard data is even more important in 2020 than open source is. Very often, public institutions have trouble transitioning to a more recent software because they can't port data from the previous software. They also find very difficult to build aggregated datasets, which most of the time are key when trying to implement global strategies. I would advocate that standard portability is imposed an all key softwares deployed in public services. It is a obviously a colossal and tricky task though...
3.
Can you exemplify what open / standard data could be used for?
Data integration is key to a great deal of problems public services have to deal with. Our healthcare system would benefit a lot from having standard data: healthcare facilities could then exchange information in a standard fashion, allowing to easily gather patient's related data. Building and spreading standards will be key to tracking greenhouse effect gas (GHG) emissions.
In France, all 500+people companies, 250+people institutions as well as 50000+people cities should fill in a detailed report about their GHG emissions every 4 years. As we speak, 70% still haven't or have provided an incomplete report. In my opinion, providing these institutions with a standard and open tool to report on their emissions would greatly help collecting and sharing this data.
TO GO FURTHER:
La plateforme Corona Status, créée par Alexis Thual et Raphaëlle Bertrand-Lalo.
Le projet Arkhn, qui vise à standardiser les données de santé en open source.