Hacker News new | ask | show | jobs
by msellout 3834 days ago
Amusingly, there's a similar database sold by the Georgia Secretary of State. Recently they accidentally included the social security number of every registered voter in the database they distributed. If I recall the news article correctly, they mailed it out on CD to 6 different customers before noticing the mistake.
2 comments

Voter registration databases are available from the Secretary of States office in each state. Anyone can request a file dump; however, some states often charge a small fee for the file. GOP and NDP use different tech stacks, but both are tied to companies (like NGP) that have built processes around obtaining and updating voter registration data yearly from each state. They then tie in historic data they have from door to door canvassing, voter turn out, and any other touch points to build a profile about the voter. Typical data points include voting history, gender, various voter level flags indicating if the voter supports causes/activism, and any other notes or voter related data.
Do they have per-voter voting history, or is it a probability based on more coarse-grained voter turnout metrics?
Per-user voting history in many (most?) states. The fact that a ballot was received from a particular voter (but not what they voted for) is often releasable under public disclosure laws.
Yes, most states track voting participation at the voter level. Some track method of voting, other just are a boolean participation flag for each election.
Unless I'm mistaken, it's impossible have a per user history.
Individual voting records aren't public in the US (secret ballot). However, whether or not an individual voted in a specific election is often public. (It may vary by state or other jurisdiction.)
In addition to the DNC voter file, there are similar lists available commercially. http://www.voterlistsonline.com/ is run by Aristotle (who tends to have more GOP clients, but is non-partisan). It has many of the consumer data fields already matched that are discussed in the article.

I would not be surprised if Aristotle wasn't one of the six customers to which Georgia sent the data. They have been collecting voter data for years; I worked there for a few months more than 15 year ago, and they already had a nationwide database mostly loaded from 9-track tape.