People
voters
We work with partners to collect and process data from each state's voter file. We then join that data with consumer data (including high quality email addresses and phone numbers), census data, and our own predictive models to provide users with a dataset that can power a variety of targeting, messaging, outreach, and modeling use cases.
Schema
| Column | Description |
|---|---|
| voterbase id | Unique voter ID |
| state voter id | State supplied voter ID, format differs from state to state |
| county voter id | County supplied voter ID, format differs by county and state |
| first name | First name in uppercase with spaces, numbers, and special characters removed. Accented characters have been replaced with non-accented versions |
| middle name | Middle name in uppercase with spaces, numbers, and special characters removed. Accented characters have been replaced with non-accented versions |
| middle init | Middle initial in uppercase |
| last name | Last name in uppercase with spaces, numbers, and special characters removed. Accented characters have been replaced with non-accented versions |
| name suffix | Name suffix in uppercase, i.e. JR, III |
| dob | Date of birth formatted as date (YYYY-MM-DD). DOB comes from voter registration data and commercial data |
| myob | Month and year of birth formatted as 6 digit integer (YYYYMM). MYOB comes from voter registration data and commercial data. Some states truncate DOB to the first of the month making MYOB better for matching in those cases |
| yob | Month and year of birth formatted as 4 digit integer (YYYY). YOB comes from voter registration data and commercial data. Some states truncate DOB to the first of the year making YOB better for matching in those cases. |
| age | Voter's age by end of current year, calculated by current_year - YOB |
| modeled age | Voter's age by end of current year, calculated by current_year - YOB, modeled if DOB is missing |
| reg date | Voter registration date formatted as date (YYYY-MM-DD): the earlier of the voter file's earliest registration date and the first election the voter is recorded voting in (dated June 1 for a primary, November 1 for a general). Some states update the registration date whenever a voter record changes, so the first recorded vote catches an older registration |
| gender | Voter gender from voter registration data. M, F, X (gender expansive), or NULL |
| ethnicity | Ethnicity, self reported on the voterfile where available, otherwise modeled. In cases where the model isn't confident, Null. AAPI, Black, Latino, Native American, White, Other (another race or multiracial) |
| ethnicity source | Ethnicity source: Voterfile, Commercial, Uncoded |
| modeled race aapi | Indigo race model, score from 0-1 with the probability that a voter is AAPI |
| modeled race black | Indigo race model, score from 0-1 with the probability that a voter is Black |
| modeled race latino | Indigo race model, score from 0-1 with the probability that a voter is Hispanic or Latino/a |
| modeled race native american | Indigo race model, score from 0-1 with the probability that a voter is Native American |
| modeled race white | Indigo race model, score from 0-1 with the probability that a voter is White |
| state fips | Registration state fips code, two digit fips as determined by the census formatted as a string |
| state | State abbreviation |
| county fips | Registration address county fips code, three digit fips as determined by the census formatted as a string |
| county | Registration address county name in uppercase |
| precinct | Registration precinct name |
| reg address | Registration address |
| reg city | Registration city name |
| reg state | Registration state abbreviation |
| reg zip | Registration address zip5 as string |
| reg zip4 | Registration address zip4 as string |
| reg lat | Registration address latitude |
| reg long | Registration address longitude |
| reg latlong accuracy | Accuracy of reg_lat and reg_long columns, ordered from most to least accurate: GeoMatch9Digit, GeoMatchRooftop, GeoMatchBuilding, RangeInterpolation, ExactMatch, AverageOfApartments, ParcelCenter, GeoMatch5Digit, KnownAlternateName, DirectionPrefixRemoved, DirectionSuffixRemoved, StreetCenter, Intersection |
| mailing address | Mailing address |
| mailing city | Mailing city name |
| mailing state | Mailing address state abbreviation |
| mailing zip | Mailing address zip5 as string |
| mailing zip4 | Mailing address zip4 as string |
| phone | Best phone number for voter, prioritizing cell phones over landlines, 9 digits formatted as a string |
| phone type | Type of phone: Cell, Landline, VOIP |
| phone confidence code | Confidence in quality of phone number High - high confidence phone number matched at the individual level Household - high confidence phone number matched at the household level Low - low confidence phone number |
| phone cell | Cell phone number, 9 digits formatted as a string |
| phone cell confidence code | Confidence in quality of phone_cell High - high confidence phone number matched at the individual level Household - high confidence phone number matched at the household level Low - low confidence phone number |
| phone landline | Landline phone number, 9 digits formatted as a string |
| phone landline confidence code | Confidence in quality of phone_landline High - high confidence phone number matched at the individual level Household - high confidence phone number matched at the household level Low - low confidence phone number |
| party | Registered party from the voterfile: Democratic, Republican, Libertarian, Green, Other, No Party Affiliation, or Unknown. Unknown where the state has no party registration or the voter's party isn't on file |
| party 3way | Registered party grouped into DEM, REP, and OTH. Null when party is Unknown |
| party source | Voterfile when party comes from registration. Modeled when TargetSmart only models a party, which leaves party Unknown |
| district congressional 2020 | Congressional district, three digits zero padded, i.e. 002, 011, 024 |
| district congressional 2010 | 2010 congressional district, three digits zero padded, i.e. 002, 011, 024 |
| district stleg upper 2020 | Upper state legislative district, state senate. For numeric districts, district names are three digits and zero padded, i.e. 003, 021, 041B. For non-numeric district names, strings are uppercase. |
| district stleg lower 2020 | Lower state legislative district, including state house or state assembly depending on the state. For numeric districts, district names are three digits and zero padded, i.e. 003, 021, 041B. For non-numeric district names, strings are uppercase. |
| census block 2020 | Census block ID, 15 digits formatted as a string |
| ts model education college graduate | TargetSmart education model. Predicts the likelihood that an individual has attained a college-level or higher education. Scores are expressed from 0-1. A higher score represents a higher probability that a person's education level is college graduate or higher. |
| ts model education high school only | TargetSmart education model. Predicts the likelihood that an individual has not attained formal education beyond high school. Scores are expressed from 0-1. A higher score represents a higher probability that a person's education level is high school or lower. |
| ts model religion catholic | TargetSmart religion score. Predicts the likelihood that a voter identifies as Catholic. Scores are expressed from 0-1. A higher score indicates a higher likelihood to identify as Catholic. |
| ts model religion evangelical | TargetSmart religion score. Predicts the likelihood that a voter identifies as Evangelical. Scores are expressed from 0-1. A higher score indicates a higher likelihood to identify as Evangelical. |
| ts model religion jewish | TargetSmart religion score. Predicts the likelihood that a voter identifies as Jewish. Scores are expressed from 0-1. A higher score indicates a higher likelihood to identify as Jewish. |
| ts model religion mormon | TargetSmart religion score. Predicts the likelihood that a voter identifies as Mormon. Scores are expressed from 0-1. A higher score indicates a higher likelihood to identify as Mormon. |
| ts model gunowner | TargetSmart gun owner score. Predicts the likelihood that an individual supports stricter gun control laws. Scores are expressed from 0-1. A higher score indicates a higher likelihood that a person supports stricter gun control laws. |
| ts model veteran | TargetSmart veteran score. Predicts the likelihood that an individual is a military veteran or an active service member. Scores are expressed from 0-1. A higher score predicts a higher likelihood that the individual is a military veteran or an active service member. |
| ts model ideology score | TargetSmart ideology score. Predicts the likelihood that an individual supports liberal ideology. Scores are expressed from 0-1. A value of 1 represents those most likely (very liberal) and 0 represents those least likely (very conservative). |
| ts model children present score | TargetSmart children present score. Predicts the likelihood that a voter lives in a household with children. Scores are expressed from 0-1. A higher score represents a higher probability that a person lives in a household with children. |
| ts model marriage | TargetSmart marriage score. Predicts the likelihood that an individual is married. Scores are expressed from 0-1. A higher score represents a higher probability that a person is married. |
| ts model income rank | TargetSmart high income score. Predicts the likelihood that an individual has an income over $100,000. Scores are expressed from 0-1. A higher score represents a higher probability that a person would self-report income greater than $100,000. |
| ts model homeowner | TargetSmart homeowner score. Predicts the likelihood that an individual owns a home. Scores are expressed from 0-1. A higher score represents a higher probability that person owns a home. |
| fec hh contribution count rep | Total number of FEC contributions to Republicans in the household |
| fec hh contribution count dem | Total number of FEC contributions to Democrats in the household |
| fec hh contribution count total | Total number of FEC contributions in the household |
| fec hh contribution pct dem | Percent of FEC contributions in the household made to Democrats |
| modeled turnout midterm primary | Indigo midterm primary turnout model, score from 0-1 with the probability that a voter will turn out to vote in a midterm primary election |
| modeled turnout midterm general | Indigo midterm general turnout model, score from 0-1 with the probability that a voter will turn out to vote in a midterm general election |
| modeled turnout presidential primary | Indigo presidential primary turnout model, score from 0-1 with the probability that a voter will turn out to vote in a presidential primary election |
| modeled turnout presidential general | Indigo presidential general turnout model, score from 0-1 with the probability that a voter will turn out to vote in a presidential general election |
| modeled race | Categorical race value based on Indigo race model. AAPI, Black, Latino, Native American, White, or NULL if model is low confidence |
| modeled dem partisanship | Indigo partisanship model, score from 0-1 with the probability that a voter indentifies as a Democrat |
| g12 reg* | 1 if voter was registered for given election, 0 if voter wasn't yet registered |
| g12 voted* | 1 if voted, 0 if didn't vote |
| g12 ballot type* | Indicated method of voting if voted. Note that not all states report type of ballot for all historic elections, a null value indicates lack of reporting. Early, Absentee, Poll Vote, Unknown. |
| p12 party* | Partisan primary voted in in given election D - Democrat, G - Green, I - Independent, L - Libertarian, N - No Party Preference, O - Other Party, R - Republican, U - Unknown |
* For general, primary, and presidential primary elections from 2012-present day, we have voted, _election_date, and _ballot_type columns for each election. The naming convention for these columns in [election stage - g/p/pp][2 digit election year - 14/22/23][column type], i.e. pp09_voted, p14_party, g22_ballot_type.