With the support of Microsoft, Bloomberg, Walmart, and Columbia DSI, the second-annual Columbia Data Science Student Challenge was a huge success. Over 220 students from varying backgrounds and schools registered and participated in the hackathon, and 30 teams survived the 24-hour marathon and presented their projects to judges from our sponsor companies. We handed out $10,000 in prizes to the four teams with the best work, but everyone in attendance made new friends, got a little better at machine learning, and slept very soundly for the next few days.
We wanted to foster a learning environment at the hackathon, where seasoned computer scientists could work side-by-side with newcomers to programming and data science. To that end, we held a number of workshops leading up to the hackathon to get everyone up-to-speed on Microsoft Azure's machine learning platform. We also held open office hours to answer questions about Azure ML and help individual participants form teams.
PRESENTATIONS from our SPONSORs
On the day of the hackathon, our sponsor companies had a chance to present to the participants. Jake Hofman (right) of Microsoft Research gave the eager hackers some practical tips on doing data science through the lens of a few of his past projects, while Bloomberg and Walmart — who each provided a dataset for hackers to analyze — explained their unique business problems and the types of solutions they were interested in seeing from the participants.
LET THE HACKING BEGIN...
After fueling everyone up on pizza and coffee, we kicked off the hacking. Teams had a choice of working with either Walmart's products dataset or Bloomberg's commodities dataset. Coding went on throughout the night, as participants cleaned, analyzed, and visualized the data. Students investigated a range of topics, from identifying cannibalistic product pairs to assessing a portfolio's risk when the market drops. The next morning, everyone submitted their work and gave a short demo to a pair of judges.
Keynote by a Facebook Data Scientist
While the judges deliberated on their favorite projects, participants were treated to a keynote speech by Ranjan Subramanian, a data scientist at Facebook. Ranjan's talk covered a variety of topics, such as how data scientists at Facebook interface with software engineers, designers, and product managers and his views on data science as a field.
And the winners are...
In the end, the judges couldn't settle on just three teams, so they announced the top three plus an honorable mention. The winning teams all managed to combined sound analysis with strong presentations, and the judges commended each of them for the work they did in such a short amount of time. A full recap of the four winners' projects can be found here.
want to get involved?
We've already started planning next year's hackathon, and we want you to be a part of it! If you're interested in participating, you can subscribe to our newsletter and like us on Facebook for updates regarding the hackathon, while potential sponsors can email us for more information.