Data on campaigns and elections is hard to find and even harder to work with. We’re here to help. At Deck, we’ve spent years building massive, complex data pipelines to gather and process data on candidate filings, election results, campaign finance records, and more. So we know how painful it can be to work with this messy, hard-to-find data. We started solving this problem in early 2021 with our launch of Hubble, our data subscription service that gives you access to everything we know about campaigns and elections from 2008 to today. The next step in solving this problem is our voterfile, Hubble Voterfile.
We are in a unique position to assemble a person table that contains the best information available about voters all in one place. Hubble Voterfile is a national voterfile that includes vote histories from 2008, census information, previous voter registration, Deck scores, TargetSmart scores and more! This also allows users to seamlessly connect person level data to robust election, geography, and results data, as well as individual level scores in Hubble which will make it easier to analyze campaigns, target voters, build clear reporting and develop custom modeling to fit your needs.
At Deck, we build a lot of models and we know that raw files can be a challenge in getting good models spun up quickly. So, we have also built an additional two versions of the Hubble Voterfile to aid you in modeling. In the modeling Hubble Voterfile, all voter information is represented as numeric features and categorical features are one-hot-encoded into binary columns. In addition, date columns are converted to features that are useful in models, i.e. year registered rather than registration date. The imputed Hubble Voterfile is built off of the modeling Hubble Voterfile where all missing data has been imputed.
How to get started
Hubble is shared as a BigQuery dataset or through a Google Cloud Storage bucket.
Find out more about Hubble and schedule a time to chat with our team at hubble.run.