[Note: I do realize that this event was nearly two months ago. I have no one to blame but the academic job market.]
On August 15 and 16, we held the first annual ASA Datathon at the D-Lab at Berkeley. Nearly 25 people came from academia, industry, and government participated during the 24-hour hack session. The datathon focused on open city data and methods, and questions surrounded issues such as gentrification, transit, and urban change.
Two of our sponsors kicked off the event by giving some useful presentations on open city data and visualization tools. Mike Rosengarten from OpenGov presented on OpenGov’s incredibly detailed and descriptive tools for exploring municipal revenues and budgets. And Matt Sundquist from plot.ly showed off the platform’s interactive interface which works across multiple programming environments.
Fueled by various elements of caffeine and great food, six teams hacked away through the night and presented their work on the 16th at the Hilton San Francisco. Our excellent panel of judges picked the three top presentations which stood out the most:
Honorable mention: Spurious Correlations
The Spurious Correlations team developed a statistical definition for gentrification and attempted to define which zip codes had been gentrified by their definition. Curious about those doing the gentrifying, they asked if artists acted as “middle gentrifiers.” While this seemed to correlate in Minneapolis, it didn’t hold for San Francisco.
Second place: Team Vélo
Team Vélo, as the name implies, was interested in bike thefts in San Francisco and crime in general. They used SFPD data to rate crime risk in each neighborhood and tried to understand which factors may be influencing crime rates, including racial diversity, income, and self-employment.
First place: Best Buddies Bus Brigade
Lastly, our first place winners asked “Does SF public transportation underserve those in low-income communities or without cars?” Using San Francisco transit data, they developed a visualization tool to investigate bus load and how this changes by location, conditional on things like car ownership.
You can check out all the presentations at the datathon’s GitHub page.
Laura Nelson, Laura Norén, and I want to give a special thanks to our sponsors: OpenGov, UC Berkeley Sociology, UW Madison Sociology, the D-Lab, SurveyGizmo, the Data Science Toolkit, Duke Network Analysis Center, plot.ly, orgtheory, Fabio Rojas, Neal Caren, and Pam Oliver.