WATTx first hackathon to advance female health tech

WE ARE WATTX

By
Laure

March 24, 2017


On the 18th and 19th of March, we held our first ever data science Hackathon around women’s health with local Berlin startup, Clue. For about 30 data enthusiasts this was a unique chance to make sense of one of the biggest sets of anonymized women’s health data and to learn about machine learning while make new connections.

For about 30 data enthusiasts this was a unique chance to make sense of one of the biggest sets of anonymized women’s health data and to learn about machine learning while make new connections.

It all started on Friday evening in our new office with a data science state of the union and a presentation on the application and future of differential privacy presented by two incredible speakers: Eleftherios Diakomichalis, VP data science at Soundcloud, and Frank McSherry, expert in differencial privay techniques. The room was packed with a very engaged audience. Recordings of the talks can be found here.

The hackathon officially kicked off on Saturday morning around a delicious breakfast and a presentation of the female menstrual cycle and the anonymized dataset that would be provided.

The privacy preserving data platform Statice (developed by us) was used during the event, enabling hackers to model the data while completely respecting the anonymity and privacy of users.

Respecting the raw data’s attributes, Statice created synthetic data accessible to the data scientists of the competition to build their algorithms on. The used software-environment and solution were then bundled in a docker image and posted on the platform. Solutions ran on the cloud-based raw data, thus enabling scoring and ranking of the solutions submitted.

Thanks to everyone for all the sharing and learning, to the speakers and to all participants and teams.

Throughout the whole weekend, 12 teams submitted more than 90 inspiring models. The first prize of 3000€ was awarded to the winning team of Adrin, Gianluca, Markus, Corrie and Marco with the best scoring model.

Congratulations!