Coopetition

Coopetition is not classic data science competition platform

In contrast to classical data science competition where the only goal is to create the best score on given metric and don't tell anyone how you did it, "coopetition" aims to build a community of people that help each other and contribute to common goal. For example, one can look at open data from different CERN detectors here.

All prediction models are open

Each prediction model that gets into the challenge is open, so everybody can get useful insights on a challenge and learn different approaches to the problem. People are rewarded for their contribution according to smart-contract attached to a "coopetition". You can get rewarded for innovation made into codebase, for posting useful insights on a challenge or for each time someone reuses your code.

"How can you achieve transparency
of contribution and rewards?"

It's done using github as a code base, so it is clear to everybody, when certain submission where made. Submission pipeline follows simple steps:
  • User commits his/her solution into their repository
  • Our service gets notified about it and picks up the solution, evaluates it and post results to the dashboard
  • Our services got notified about it

If you want to participate in coopetition, it is as simple as it could be:

  • Go to a Competition list and find an active coopetition that you want to participate.
  • Register in codalab, with the same email you have on github.
  • Fork baseline repository and follow this instruction to integrate your repository into the pipeline.
  • If you want to see solutions of others checkout leaderboard of a competition to find a submission which might inspire you. Good luck!