Excited to share the latest DrivenData competition, which features a tricky data problem of bootstrapping a good prediction off of limited data in a cold-start scenario.

Should be an interesting challenge, with a unique structure and uncommon dataset!