Unlocking Novel Strategic Campaign Insights for 1.56M Voters
With the 2024 election impending, Resilient Strategies AZ wanted to utilize a unique dataset it knew it had that displayed individual votes within ballots and their geographic attributes, however it had run into consistent problems attempting to extract the massive amounts of data from these files. The company wanted to utilize the data from 2022 to help optimize their strategies for their candidates, using visuals to understand how voters who voted for one particular candidate, voted in their down ballot races.
Challenges
How can we utilize the data from those files to understand how voters for one candidate in one election vote in the other elections on their ballot?
How can we create a structured dataset out of 7,000+ files of semi-structured data?
Key Results
Loaded, extracted, and modeled 7,000+ files, or 70+ GB, of semi structured data for analytics usage within a relational cloud database
Created and published dashboard that provided end users down ballot insights for over 1.56 million voters for 247 separate elections within Maricopa County in 2022
“It feels like we discovered a completely new, smarter way of analyzing elections.”
— Eric C. , CEO of Resilient Strategies
Value to Business
Resilient Strategies AZ has a “one-of-a-kind” tool that allows them to pair insights with previously utilized messaging for candidates and contests to understand how they can win in precincts and districts where, on the surface, it appears that their candidate might not do well.
We utilized cloud storage and database infrastructure to extract the semi-structured data from the files and then unpack that data into a relational database with a data model. Then, we analyzed that data and created a reporting structure that we then loaded into a data visualization tool to create the visuals that allowed the end user to filter the data for the down ballot contests based on how a voter voted in a primary contest.