Housing Condition Scores (Census tract, ZCTA, and H3 Grid)

The Housing Condition Scores dataset provides aggregated housing condition scores at the census tract, ZIP Code Tabulation Area (ZCTA), and H3 hexagonal grid (resolution 6) levels across the United States. A machine learning model was used to impute missing housing conditions in the original Warren Group data, which classifies each property as Excellent, Good, Average, Fair, Poor, or Unsound. Individual property scores were then combined into each spatial unit using a weighted-average method ranging from 1 (lowest) to 6 (highest). We acknowledge access to the Warren Group’s property data through a data use agreement between Florida State University and Dewey.