Research Methods:

• The 1959 data was digitized from a bound copy of the City of Phoenix Advanced Planning Task Force’s 1959 study Land Use of the Phoenix Urban Area1 based on a map of proposed 1960 Census tracts and data tables detailing land-use by acreage by tract for the categories “Single-Family”, “Two-Family”, “Multifamily”, “Commerce”, “Industry”, “Streets and Alleys”, “Public and Semipublic”, “Agriculture”, and “Canals and Vacant” (43-46).
• Microsoft OneNote was used to convert the scanned 1959 land-use tables to text. The resulting tables were then exported to Microsoft Excel and corrected manually to match the original scan.
• The scanned tracts map was georeferenced in ESRI ArcGIS and compared to a 1960 Census Tracts shapefile available from the National Historic GIS database (NHGIS). Boundary mismatch was corrected by consolidating some tracts (denoted by hyphens in the “TRACT_NUM” field). The data tables were aggregated accordingly and joined to tracts in ArcGIS.
• The 2012 time point was generated using 2012 Maricopa County Assessor parcel data with land-use classified based on the property use code (“PUC”) field (see “Phoenix Land-Use and Zoning by Urban Villages” dataset, also available through CAP; land-use for this dataset was classified from an earlier April 2013 version). Land-use was then reclassified to match the categories present in the 1959 Land Use of the Phoenix Urban Area, using the classification scheme outlined in Harland Bartholomew’s 1955 Land uses in American cities2 as a reference.
• Classified parcels were intersected by 1960 tract boundaries to split large parcels falling in more than one tract. The acreage of each parcel (“Area_Ac”) was calculated, and parcels were aggregated to census tract boundaries by category using a Spatial Join.
• For the 1959 data, “Single-Family” and “Two-Family” residential were consolidated into a single “SFR” class, which was already classified for the 2012 data. “Streets and Alleys” were omitted from the aggregated land total for each tract, as the 2012 data already accounts for their absence.
• Since the 2012 parcel data also omits canals, the acreage of canals in each tract for 1959 was subtracted from the “Canals and Vacant” class to generate a Vacant-only (“VAC”) class. This was performed by buffering a 1961 Canals shapefile digitized from historical Salt River Project canal system maps available at the ASU Libraries Map Collection. For smaller lateral canals, 20-foot flat buffer estimates of the canal and right-of-way were calculated; for major canals (Arizona Canal, Grand Canal, Western Canal west to Elliot Rd), 45-foot flat buffer estimates of the canal and right-of-way were calculated. The buffer polygons representing canals were then intersected by tract boundaries, and their acreages were subtracted from the initial “Canals and Vacant” total to generate the “Vacant” class. Only one tract (7-58) resulted in negative area, which was reset to 0 acres under the assumptions that a) all vacant land originally classified for 1959 was a canal or right-of-way and b) that the 1961 canals shapefile under-accounts for the length of canals present in 1959 (the tract is largely agricultural in 1959).
• For each time point, a “Land_Total” class summing the acreages of each category by tract, excluding canals/streets/alleys, was calculated. Finally, the shares of each land-use category were calculated based on the “Land_Total” field.
• 1Advance Planning Task Force, Phoenix and Maricopa County, Arizona. (1959). Land use of the Phoenix urban area: a study basic to long-range planning. Phoenix, AZ: City of Phoenix.
• 2Bartholomew, Harland. (1955). Land uses in American cities. Cambridge: Harvard University Press.