Below is a list of the steps we took to prepare the data. Those that are county specific are marked.
Restructuring
- Select variables of interest
- Choose Residential only properties based on Recoded Land.Use (e.g. removing hotels).
- Remove vacant land (NA improvement value)
- Kept mobile home even though no improvement value in James City only)
- Remove parking lots (NA land value)
- Remove Common Area (land use code – James City only)
Cleaning
- For parcels with condo CoreLogic Land Use code but apartment County Land Use, change property type to match county code (Arlington only)
- For parcel that was classified as single family but found to be multifamily though search, make into multifamily.
- Cleaned any invalid entries found.
- 43 bedrooms for a parcel in Arlington County for 2009, 2010
- Misspelling of Williamsburg across all years for James City County,
- Recode Absentee.Owner.Status into clearer factor names
- Single family units are listed as having 0 units. These were changed to 1.
Transformation
- Recode Land.Use into different residential housing types.
- Recode Number of Units in Building to match ACS categories using Land.Use Code and Number of Units
- Create own Census tract and block group IDs based on lat and lon
- Recode year built into ACS categories
- Recode number of bedrooms into ACS categories
- No way to code 0 bedrooms
- Create constant 2013 equivalents for Improvement value, Land Value, Total Value, and Taxes
- Place these into appropriate ACS categories