w2 hourly status

WarningStaging

Built from the latest training run on 2026-09-30 17:45 UTC, not from a released model set. The released documentation is at https://matter.indigo.engineering/docs/.

Overview

This model predicts whether a voter is employed as a W-2 hourly worker.

Trained on survey data, it leverages age, education signals, political engagement, and demographic features to identify likely hourly wage earners.

Organizations can use these scores to target outreach around minimum wage, overtime protections, scheduling fairness, and workplace benefits.

  • approach: dnn classifier
  • training rows: 8,154
  • last updated: 2026-09-02 19:09:13

Validation

metric value
area under ROC curve 0.73
area under precision-recall curve 0.26
accuracy 0.58
precision 0.21
recall 0.83
f1 score 0.34
mean log loss 0.85
gini coefficient 0.45
brier skill score 0.09

Lift Chart

Each bar shows how much more common the outcome is within a score decile than in the population overall. Bars taller than the dashed line (lift of 1) mean the model’s top-ranked groups concentrate the outcome; a steady decline from decile 1 to 10 indicates the model is ranking effectively.

Calibration Plot

Bars of similar height within each decile indicate well-calibrated predictions. Systematic differences reveal where the model over- or under-predicts.

Cumulative Gains

The further the model curve bows above the diagonal, the better the model concentrates the outcome in its top-ranked predictions.

Decile Analysis Table

decile count mean actual mean predicted mae lift
10 199 0.32 0.34 0.02 2.49
9 198 0.21 0.24 0.03 1.67
8 199 0.18 0.18 0.001 1.39
7 198 0.15 0.13 0.02 1.19
6 199 0.20 0.10 0.10 1.58
5 198 0.05 0.08 0.02 0.40
4 198 0.07 0.06 0.003 0.52
3 199 0.04 0.06 0.02 0.28
2 198 0.03 0.05 0.02 0.24
1 199 0.03 0.04 0.01 0.24

Architecture

This model is a deep neural network classifier trained with BigQuery ML.

Hyperparameter configuration

  • max iterations: 20
  • learn rate: 0.001
  • l1 regularization: 0
  • l2 regularization: 0
  • min relative progress: 0.01
  • warm start: FALSE
  • early stop: TRUE
  • input label columns: employment_status_w2_hourly
  • data split method: AUTO_SPLIT
  • hidden units: 64, 32, 16
  • batch size: 16
  • dropout: 0.0276
  • num trials: 18
  • max parallel trials: 5
  • hparam tuning objectives: ROC_AUC
  • enable global explain: FALSE
  • tf version: 1.15
  • auto class weights: FALSE
  • activation fn: relu
  • optimizer: adagrad

Features

Importance

Sourcing

OpenSecrets

  • donations total (count)

Revelio Labs

  • avg salary differential
  • broad job category
  • current position tenure (days)
  • job category confidence

State voter files (via TargetSmart)

  • age
  • registered republican

TargetSmart

  • activist score
  • cannabis legalization support score
  • children present score
  • double negative gop score
  • ideology score
  • income rank
  • paid leave support score
  • race: latino
  • race: native american
  • urbanicity score
  • working class score

U.S. Census American Community Survey

  • block group % income $200k+

Implementation

Prediction distribution

statistic value
mean 0.25
std dev 0.15
min 0.05
25th 0.11
median 0.25
75th 0.37
max 0.93

Segments

segment lowest score highest score share of records
Most likely 0.37 0.93 25.5%
Somewhat likely 0.26 0.36 24.0%
Least likely 0.05 0.25 50.4%

External validity

Our model indicates that 25% of people with scored records are employed as a W-2 hourly worker. BLS CPS 2024 places W-2 hourly workers at roughly 55–60% of employed workers, or ~35% of all U.S. adults.