This is a project to predict Bixi numbers, using the historical data that Julia/Kamal/Fabrice have been scraping (Fabrice is not involved; he just has data)
The idea is to make a webapp where people can go and
- Get projections of how many bikes there will be at a given station for the rest of the day
- Find out when they need to leave work to get a bike near them.
Right now there is no code and we are working on
- an API for communicating between the predictions engine and the frontend
- a data store to for the training and the predictions engine to interact with
See the issues and the wiki for discussions. If you're interested, please edit/comment on the issues!