DSI's Center for Health Analytics Launches COVID-19 Data Challenge
The COVID-19 pandemic needs no introduction, especially not to Columbians, who are at the front lines of the fight. It is clear that effectively managing this situation critically depends on extracting valuable predictions and insights from the rapidly accumulating data about this infection. The Center for Health Analytics at DSI has launched a short-term data challenge, aimed at crowdsourcing the skills in our community towards both immediate impact as well as longer term efforts. See details below.
Itsik Pe’er and Lena Mamykina
Co-Chairs, Center for Health Analytics
This COVID-19 Data Challenge is open to members of the Columbia University community. Each team of two to four members is allowed one non-affiliate member who is not the captain.
Awards for Tracks 1 and 2:
1st place: up to $2400 ($600 per team member)
2nd place: up to $1600 ($400 per team member)
Apply to join the organizing team as a volunteer, judge, hire, or data contributor.
Predict values of the cumulative number of confirmed cases, recovered cases and deaths for May 1, 2020.
Submission format for Track 1:
Upload three .csv files (for the three target numbers, same format as in, e.g., https://github.com/CSSEGISandData/COVID-19/blob/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_global.csv, with a column labeled 5/1/20 with positive integer predicted values. Teams will be given a folder to upload files.
Link to the team’s repo with code generating aforementioned files
Evaluation for Track 1:
Sum, over all entries that are ≥100 on Friday, April 24, 2020, 11:59 p.m. EDT, of the squared difference between log(prediction) and log(truth).
Allowed data and methods for Track 1:
Any legal data and methods are allowed, as long as the code is provided.
Open-ended creative analysis of COVID-19 data, but judged on interpretability and creativity, rather than accuracy.
Submission format for Track 2:
Six non-animated slides, per the template in the folder provided to teams, for a six-minute presentation
Link to the team’s repo with code generating reported results
Evaluation for Track 2:
Final Project Submission Deadline: Friday, April 24, 2020, 11:59 p.m. EDT