TurfCutter: When a friend and I volunteered to canvass for Democrats in 2006, we were appalled that the turf-cutting process still involved paper maps, scissors, and highlighters. The next summer I developed TurfCutterGM, a Google Maps-based canvassing app. VAN (now NGPVAN) licensed the product for $1, thus making it available to all Democratic campaigns. Problem solved.
Election Day Resource Calculator: To help fulfill Obama’s re-election night promise of “we have to fix that” in reference to long lines at polling places, I developed a resource calculator for election administrators. I tried to make it as simple as possible so that county officials didn’t need to already have access to complex stats (e.g., rate of voter arrival) to use it. The calculator is perhaps too simplistic (e.g., assuming presidential turnout). I plan an update in the summer of 2014 to help with the midterms. The calculator was featured on the Presidential Commission on Election Administration’s website (Bauer-Ginsberg), and received a bit of press to boot.
Open Civic Data Division Identifiers: In coordination with the Sunlight foundation and the Google civic innovation team (i.e., as I consulted for them), we launched unique, machine- and human-readable identifiers for nearly every political jurisdiction in the US. And, as of this writing, the project has expanded into five other countries. This blog post is a great overview; in brief, the identifiers let civic organizations easily match and share their data. Conveniently for organizations who want to adopt these identifiers (as Open Elections already has), Google returns these identifiers in its Civic Information API.
Text Messaging Experiment: Allison Dale and I, with help from several awesome organizations, conducted the first large-scale text messaging voter mobilization experiment. From our experiences, we developed the Noticeable Reminder theory of mobilization, which both complements and contrasts with the dominant Social Occasion theory. Other organizations have since replicated our experiments and provided further support for the Noticeable Reminder theory.
Voting Information Project: Where do I vote? Who is on my ballot? These questions are surprisingly hard to answer in the United States. I was the founding Domain Specialist of the Voting Information Project (video), a joint venture by Pew and Google (later joined by NOI, Microsoft, and Engage), which disseminates election data to anyone who wants to broadcast it (e.g., CNN, Facebook, campaigns, newspapers, random tumblr).
Persuasion Microtargeting: Kosuke Imai and I wrote a paper on heterogeneous treatment effects of randomized political voter contact experiments, which took off much more than I expected considering how many words I just had to use to introduce the concept. The method, which I have since refined and improved with the tremendous help of the Analyst Institute, identifies voters who are actually persuaded by political messaging (rather than guessing if voters will be persuaded by their answers to other questions).
All for Good: Google developed All For Good (now owned by the Points of Light Institute) to make searching for volunteer opportunities as easy as possible. I wrote the XML spec, but other people smarter than I handled the engineering.
Dissertation: It may be a stretch to classify this dissertation as a success, but the celebration certainly was…into the fountain! (True, my cell phone is in my left pocket in this picture, but after some rice-drying it worked again.)
2048 Algorithm: I spent a day writing an algorithm to beat the addicting game 2048. Not very substantive, but satisfying.
Acting Intelligently: A Brief History of Political Targeting: My friend Nathaniel Pearlman wanted to produce a book that went beyond standard media treatment of technology and politics. Rather, he strove to provide a realistic portrayal of the benefits and limitations of tech in campaigns, and I was very happy to help in the endeavor. I have great respect for many of my fellow contributors; this edited volume is worth a read.
Mobilizing the Mobiles: Text Messaging and Turnout: In this edited volume, Allison Dale and I write about our text messaging study from a practioner’s point of view. There are lots of interesting findings from our post-treatment survey that didn’t make it into our academic article.
Carnival Booth: An Algorithm for Defeating the Computer-Assisted Passenger Screening System: My good friend, Samidh, and I demonstrated why one of the government’s first attempt attempts at post-9/11 airline security was doomed to fail. This study receive a fair amount of attention and when then-Sec. Ridge shut down the program in 2004, he cited the procedure’s ineffectiveness.
“Less” Successful Projects
ROCI: My attempt to move custom organizer regions and progress-to-goal maps from spreadsheets and paper maps (respectively) to a Flash application. Turned out that not many people wanted this. Oh well, I thought it was cool and at least it worked. (FYI: pronounced ROCKY, stands for Return on Campaign Investment.)
I’m zero for one in terms of start-ups: I probably should have learned my lesson from the failure of ROCI, but I didn’t. Oops.