The Bayes Impact Hackathon
bayesimpact.org
bayesimpact.org
If I'm reading this correctly, the grand prize is having your project taken from you and losing attribution? And the only reward the team gets in return after hours of work is breakfast, lunch, and dinner?
If anyone could clarify the nature of the grand prize, I would appreciate it.
The question actually gets to something we believe in very strongly: real impact on social problems takes focus and commitment. That's why we run a full-time data science fellowship, rather than have volunteers. Many well intentioned skill-based volunteer projects end without results, mainly because at some point (usually when the work get boring or hits an obstacle) volunteers lose momentum. Having full-time fellows allows us to bring really smart people to a great idea and have them fully dedicated to seeing things through, from scoping to implementation and maintenance.
Most hackathons projects (especially the ones for social good) usually don't get continued beyond the event. We've seen this to be true, both from our own experiences at hackathons and from others. We don't want that to happen at ours, especially with such incredible civic and nonprofit organization partners. That's why we're using our own resources to dedicate 2 of our fellows, full-time, in continuing on the winning hack. We want to ensure that the great work done for social impact doesn't get lost post-hackathon.
Of course, if the winning team wishes to continue on their for-good project with our project partners, we'd love that and would do nothing to stop it. I guarantee you that we're not trying to take attribution from participants, but rather want to give teams the opportunity to see their weekend hack come to life.
The hackathon will feature fairly exclusive datasets from our project partners, so it'll be tough for folks to do a bunch of work before the hackathon.
The actual hackathon will be groups working on problem prompts and datasets provided to us by our project partners and likely judged on creativity, practicality, technical brilliance, and use of data. A sample dataset and problem might be looking at a text database of escort ads to pull out pricing data (which is heavily masked in slang) to build a dynamic demand model. The goal would be to help Thorn (one of our partners) and investigators identify movement of child sex traffickers as it happens.