Overview

Kensho's Team Impact is excited to partner with the American Voter Project (the non-profit that runs the Ohio Voter Project) to make this dataset on Georgia voters available via Kaggle.

This dataset has two main components. The first is statewide Georgia voter lists for October, November, and December provided by the American Voter Project and originally sourced from the Georgia Secretary of State. The second is cartographic boundary files from the US Census.

Starter Notebook

Jump right in with a starter notebook that demonstrates reading the data, creating maps, and aggregating voter data.

https://www.kaggle.com/gabrielaltay/georgia-voter-list-starter

Voter File Descriptions

Voter files contain one row per person, are provided for October, November, and December of 2000, and use the following naming convention,

tbl_prod_GABUYYYYMM_sample.csv

The samples are defined as,

• all: all voters in the file provided by the secretary of state of Georgia for a given month
• dropped_records: voters that were in the all sample last month but are not in the all sample this month
• new_records: voters that are in the all sample this month but were not in the all sample last month
• address_change: voters with address info that changed from last month
• name_change: voters with name info that changed from last month
• voter_in_inactive: voters with voter_status = I in the all sample for this month
• voter_status_change: voters with voter_status that changed from last month

Due to privacy concerns we have removed names and addresses (except city, zipcode, and county) from the voter files.

Geographic File Descriptions

The geographic data we collected consists of geojson files that describe cartographic boundaries in the US. We obtained shapefiles from this website and converted them to geojson using geopandas. We follow the naming convention used for the census shape files,

cb_2019_us_entity_rr.geojson where,

entity = the geographic entity rr = resolution level (we use the 20m = 1:20,000,000 and 500k = 1:500,000 scale files)

Specifically, we include the following geographic entities,

cbsa: metropolitan / micropolitan statistical area cd116: congressional district (116th congress) county: county csa: combined statistical area division: national division (subdivisions of regions) nation: national outline region: national region (northeast, southeast, midwest, west) state: state and equivalent zcta510: 5-digit ZIP code tabulation area (Census 2010)

Usage Restrictions

A quote from the Georgia Secretary of State dataset website,

>The Statewide Voter List is an electronic file that includes the date last voted for each registered voter in the state of Georgia.

>By law, voter registration lists are available to the public and contain the following information: voter name, residential address, mailing address if different, race, gender, registration date and last voting date. The Statewide Voter List does not include telephone numbers, date of birth, Social Security number or Drivers License number. The Statewide Voter List includes Active and Inactive Voters.

>Normal production time is 1-2 weeks upon receipt of order. The Statewide Voter List file will be provided to you electronically.

>The pricing is set by the Secretary of State office. This data may not be used by any person for commercial purposes. O.C.G.A. § 21-2-225 ( c )

>In accordance with O.C.G.A. § 21-2-601, any person who uses the list of electors provided for in O.C.G.A. § 21-2-225 for commercial purposes shall be guilty of a misdemeanor.

Acknowledgements

We would like to thank Steve Tingley-Hock in general for his years of work on behalf of voters and specifically for sharing this data. You can learn more about his work at the following links,

• (VICE) This Man Exposed 40K Voters Purged by Mistake
• (Wired) One IT Guy's Spreadsheet-Fueled Race to Restore Voting Rights
• (New York Times) Ohio Was Set to Purge 235,000 Voters. It Was Wrong About 20%