I attended my first hackathon today and my team, save the worRld, won best model! Greg Faletto, Shruhi Desai, Sam Park, Faizan Haque - I couldn’t have asked for a better team, congrats! To the OC R Users Group team, John Peach and the Amazon EC2 resources that he provided, to all the volunteers, and to UC Irvine’s Paul Merage School of Business for hosting, thanks for putting on an amazing event. It was super educational and the perfect environment to develop coding skills (and networking too)!
In 24 hours, my team used R and Tableau to explore, analyze, and create predictive models using California publicly available water source data, health data, census information, etc. We were very time constrained (hence our findings were preliminary more than scientific), but with fairly high accuracy we developed a model that can predict the percentage of poor-health residents from drinking water contaminates (we analyzed chemical levels of arsenic, nitrates, and uranium).
We also explored how these contaminates affect low birthweight and child poverty percentages at the county level here in California. Greg wrote an overview of our methodology and model on his blog (link below) and I published some of the exploratory visualizations (sample dashboard below) to my Tableau Public profile (these interactive Tableau visualizations are not mobile friendly!). All around great times and I’m already looking forward to the next one! We plan on refining our code and then we’ll make a GitHub repository public for anyone who wants to replicate our methodology and models.