Simply put, vote similarity compares each legislator’s voting history next to every single other legislator. These stats between each legislator are a result of collecting high quality data sets and then asking the right questions. In this case we’re dealing with a
n x n (per legislative session and chamber) question, this means in a chamber with 110 legislators we’ll generate 12,100 data points. As we compare each legislator we exclude votes where either of the two legislators being compared were absent or excused, this is to only base our results on actual opinions expressed.
To query this efficiently we’re using a serverless computing service on Amazon Web Services called Athena, for the technically interested this is a serverless data querying service running on top of Presto (by Facebook) which allows us to run SQL queries on big data sets affordable, scalable and fast. If you’re interested on reading more about how we’re doing this, Statehill co-founder Karl Roos wrote an article on Medium about it.
The vote similarity feature is located directly on legislator pages. By default the top three most similar legislators are displayed—the complete list of legislators relative to a particular person is also available.
Bottom line, in the new political-tech era, everything is connected. Aggregating and presenting information like vote similarity helps you connect the dots to gain insight towards the bigger legislative picture. For any questions, comments, or to learn more about how Statehill tools and data can add value to your legislative operation, get in touch at firstname.lastname@example.org.