During an event hosted by the UC Davis Data Science Initiative a partner and I chose to visualize Proposition 11, which deregulated emergency medical personnel and their break time. Our research uncovered that the proposition was entirely funded by American Medical Response (AMR), the largest private ambulance service in the area. I scraped several tweets regarding Proposition 11, including those which used clearly aligned tags like #YesOn11 or #NoOn11. Using the aligned tweets, I created a neural net classifier to sort the tweets that were not clearly aligned (like #Prop11) into the support or opposition camps. Analysis demonstrated that positively aligned tweets came almost exclusively from AMR sources. Read more about this project here.
While the original presentation was not recorded, my partner and I gave a presentation two years later at the next ballot proposition challenge. You can watch that short presentation below.