Initiated by ETH Zürich and partner organisations
What is "Epidemic Datathon"?
It is a collective open source real-time forecasting challenge that aims at joining forces to develop real-time and large-scale epidemic forecasting models.
After the initial spread of SARS-CoV-2 in Hubei (China), the World Health Organization recently declared the corona virus disease COVID-19 a pandemic. The goal of this project is to use publicly available data to accelerate scientific innovation in modeling and forecasting the evolution of COVID-19 cases in different countries as well as to evaluate possible response measures. Quantifying the forecast accuracy and uncertainty in real time is used as a benchmark for epidemic forecasting models and hopefully can provide additional insights into the COVID-19 outbreak dynamics. Finally, we plan to contribute to the global open-source scene by releasing real-time epidemic forecasting models.
Are mechanistic epidemic models able to make good predictions or do purely data-driven approaches outperform standard epidemiological frameworks? Better performances of data-driven models that incorporate various datasets may help to determine missing features in standard epidemic models. Large deviations in the predictive accuracy of standard epidemic models can indicate wrongly estimated disease parameters and containment strategies.
We encourage participants to get inspired by the state-of-the-art research in epidemiology, network science, statistics, data science, and other fields (see references for further details) and make scientific contributions. Please also review our disclaimer and ethics section.
Who can join?
Everyone can join and contribute in various ways: (i) register as a developer (individual or with a team) of a real-time epidemic forecasting model, (ii) register and monitor scientific developments (see our disclaimer section), or (iii) share the news about this event and help us reach more contributors.
Official starting date: Monday, March 30, 6 pm CET, 2020
End date: to be announced (until the outbreak stops)
Join datathon and develop
Official datathon for the ETH Zurich class: 851-0585-38L Data Science in Techno-Socio-Economic Systems & EU SoBigData++ datathon.
Nino Antulov-Fantulin (Computational Social Science, ETH)
Dirk Helbing (Computational Social Science, ETH)
Lucas Böttcher (Computational Medicine, UCLA &
Institute for Theoretical Physics, ETH)
Zhang Ce (DS3-LAB, Computer Science, ETH)
David Dao (DS3-LAB, Computer Science, ETH)
Hans Gersbach (Macroeconomics: Innovation and Policy, ETH)
Christopher Dye (Oxford Martin School, University of Oxford)
Dino Pedreschi (University of Pisa, Italy)
Fabrizio Lillo (Department of Mathematics, University of Bologna, Italy)
Mile Sikic (A*STAR Genome Institute of Singapore, FER University of Zagreb)
Petter N. Kolm (NYU Courant, USA)