The weekend before last, I had the pleasure of participating in the inaugural EMC Data Science Global Hackathon. The event, orchestrated by kaggle as part of Big Data Week, involved teams around the world competing to build better, more accurate predictive models of metropolitan air pollution. Air pollution threatens millions of people around the globe suffering from asthma and other respiratory diseases each day, and predictive models like these can provide an early warning system to alert the public regarding dangerous levels of pollutants on an hourly basis.
There was a global competition, in which teams from London, New York, Boston, Chicago, San Francisco, Melbourne, Canberra, Sydney, and Turku, Finland (as well as independent entrants from around the web), battled for 24 hours straight to top the leaderboard by building the most accurate prediction models. The NYC event also included a Data Visualization competition: who could transform the competition data into the most effective and impactful visual form.
My team focused on this Data Visualization competition. Lo and behold: we won! Team CornerOffice (comprised of Frederic de Sibert, Gregory Gorin, and myself) “out-visualized” the other teams competing in the 28th floor offices at Bloomberg to take the top prize. Our submission leveraged a data analysis and visualization tool called Tableau to create customized, interactive dashboards providing insight (and some animation) around the time, location, season, weather, wind, and other factors affecting pollution levels in the Chicago metro area.
Several other competitors boasted impressive visualizations as well – including a simulated public website for accessing pollution prediction information, and another interactive dashboard showing how changing the ranges of any attribute in the data set affected the distribution of all other factors. Thus we were in good company as the esteemed panel of judges – Cathy O’Neil (Intent Media, Mathbabe.org), Chris Wiggins (Columbia University, HackNY), and Jake Porway (NYTimes, DataKind) – chose our entry for the win.
Having unfortunately elected to take a cruise to the Bahamas last December (ok I’ll admit, probably not the most accurate use of the word “unfortunately”), I missed last year’s annual TheLadders Hackathon, and thus last Saturday marked my first time participating in this kind of event. I was glad to get the W, but even more so to learn a great deal, meet a bunch of great people (especially my teammates Greg and Freddie), and finally see what all this “hackathon fuss” was about. Many thanks must go as well to Matt Turck and Shivon Zilis of Bloomberg for providing the impressive venue and supporting the NYC arm of this global competition. Though this was my first hackathon, it certainly won’t be my last!
Noah Goldenberg is “The Data Guy” (aka Research Analyst) at TheLadders. He’s a New Jersey native, but also spent time as a Virginian and occasionally still says “y’all”. When he’s not wrangling data or winning hackathons, he likes to take it easy and run marathons.