Protecting the Vote

Common Cause × Illumina Consulting Group, 2020

I had a front-row seat to one of the most consequential and chaotic elections in US history. Common Cause was on a mission to predict and respond to voting interference and difficulties in the 2020 general election. Illumina had the tools and expertise they needed; I bridged the gap. In all, our work helped Common Cause assist more than 400 voters and counter disinformation narratives as they emerged.

Working cross-functionally with Illumina’s engineers and analysts, I created rulesets on our proprietary real-time streaming analytics platform LUX to detect voting abnormalities during the early voting period. I taught Common Cause’s Voting & Elections team how to use LUX to create and modify rules, and together we manually evaluated the outputs to fine-tune these rules ahead of election day. LUX’s alerts allowed Common Cause to identify and respond to pertinent situations in real-time.

I authored a final report for Common Cause summarizing our collaboration. You can read it here.

Encore: LUX Voter Watch 2022

My role was, first and foremost, to read the news, but with a twist: I wasn’t looking for what happened yesterday or today. Instead, I needed to analyze these events to look ahead to what was likely to happen in the coming weeks and months. Each day, I updated a spreadsheet of topics, sorting between relevant and irrelevant stories. I also looked for hints of stories to come (seeing the abortion case on the Supreme Court docket, for instance, was certainly prescient). I further grouped these into subtopics and predicted the length of each news cycle. Once we had a sense of the issues, I created rule sets in Illumina’s data streaming tool, LUX, to detect social media posts about each issue. LUX’s machine-learning algorithms then sorted each post based on gender, political affiliation, sentiment, and location. These algorithms can additionally infer how the post authors feel about issues in regard to specific candidates. I continually reviewed LUX’s outputs, refining the rules as necessary to increase their accuracy as well as adding new issues as they arose.

The end result was a beautiful visualization of issue and candidate tracking in real-time.