Tal Salmona, Chief Architect
Last week, 10 teams gathered at our Tel Aviv office to join our first ever public Data Hackathon.
In this Hackathon, teams had a chance to prove their big data and analysis skills by solving a real-world problem over huge real-world dataset.
The Problem: Teams were asked to predict the “money in the funnel” problem. Given a path of historical user events (clicks and conversions), one should try to predict the future revenue potential that is expected for the user.
The Goal: The winning team was the one who was able to most accurately predict the value of future conversions.
The Data Set: 50 Terabytes (TB) of anonymized clicks and conversions event streams.
The Hardware: Amazon Web Services (AWS) helped facilitate this Hackathon by providing the teams with accounts (and budget) to run their analysis on the AWS cloud. They also provided people who helped the teams make the best use of the tools available on the AWS cloud.
The teams worked around the clock for more than 48 hours to come up with their algorithms, code it, test it and run it on the full data set. It was really interesting to see the different approaches they took and the passion they had towards solving this problem.
Most of the teams made it to the finish line, some were way off, some were really close. The winning team (Yizhar Gilboa, Tal Barda, Inbal Valigany, Iddan Golomb) received a nice deep-learning programmable robot and more importantly, eternal fame and recognition!
We learned that it’s not enough to be a good algorithm researcher and it’s not enough to be a master of big data tools – you need a team that can do both.
We are already planning our next Hackathon event. We have some really cool ideas of what will be the next challenge we’ll ask the community to solve. Stay tuned!
Here are some photos from the event:
Algo solving needs a lot of coffee
Brain power in action
No food, No code
First place prize – what a cool robot
Tired but happy – Mentors wrapping up the event and going home