You will receive data of the form (admin_code, date, Ebola case count) and (admin_code, covariate value). Your objectives are the following:
1) Train a model / mine associations between the time-series and the covariates for the case data though Oct. 1.
2) Test that model / those associations on data after Oct. 1 (honor system for not using this data in training).
3) Tell us (a) how well your model performed (prediction accuracy) as well as (b) the key factors in your model. Discuss anything you think might assist in making decisions to contain the outbreak (example findings might be that certain factors seem to matter a lot or certain regions are likely to be hard hit in the near future). Also discuss limitations due to the models and the data.
We encourage you to collaborate with other teams. The discussion forums are there to assist with this; there will also be several crisis and ebola experts participating remotely on this. To make sure that questions and insights get seen by the right people, please use the following in your posts:
1) INSIGHT: [ description ] to discover findings
2) DATA ISSUE: [ description ] to discuss things related to the data
3) MODELING ISSUE: [ description ] to discuss things related to the data
4) NEED: [ description ] to discuss a need
As always with forums, always check if there's already a relevant thread before starting a new one.