Data for: URBANE: Urban Reasoning and Block Adjustment via Natural-Language Editing

The dataset used in URBANE: Urban Reasoning and Block Adjustment via Natural-Language Editing. It consists of 12 .pkl files, one per U.S. city. Each file contains a list of dictionaries, each dictionary representing an individual city block from that city. The data distribution is as follows: Houston, TX: 17,253 blocks Chicago, IL: 14,291 blocks New York, NY: 13,017 blocks Phoenix, AZ: 12,651 blocks San Antonio, TX: 11,912 blocks Dallas, TX: 9,233 blocks Philadelphia, PA: 8,798 blocks Jacksonville, FL: 8,771 blocks Columbus, OH: 5,938 blocks Austin, TX: 5,528 blocks San Jose, CA: 3,298 blocks Miami, FL: 1,502 blocks Each dictionary has the following features: Geographic/Administrative Feature Description Range/Type SF State String CF County String city City String Environmental & Climate Risk Feature Description Range/Type DSF_PFS Diesel particulate matter exposure [0, 1] EALR_PFS Expected agricultural loss rate [0, 1] EBLR_PFS Expected building loss rate [0, 1] FLD_PFS Share of properties at risk of flood in 30 years [0, 1] WFR_PFS Share of properties at risk of fire in 30 years [0, 1] IS_PFS Impervious surface or cropland coverage [0, 1] Socioeconomic Indicators Feature Description Range/Type HBF_PFS Household financial burden [0, 1] LMI_PFS Low median household income (percent of area median) [0, 1] HSEF Percent individuals age 25+ with less than high school degree [0, 1] P100_PFS People with incomes below the poverty level [0, 1] FPL200S Is low income {0, 1} LHE Low high school education {0, 1} UF_PFS Unemployment [0, 1] Housing & Infrastructure Feature Description Range/Type LPF_PFS Percent pre-1960s housing [0, 1] KP_PFS Share of homes with no kitchen or indoor plumbing [0, 1] TF_PFS Traffic proximity and volume [0, 1] TSDF_PFS Proximity to hazardous waste sites Numeric Health & Demographics Feature Description Range/Type LLEF_PFS Low life expectancy [0, 1] EPLR_PFS Expected population loss rate [0, 1] TPF Total population [0, 1] AGE_10 Percent age under 10 [0, 1] AGE_MIDDLE Percent age 10 to 64 [0, 1] AGE_OLD Percent age over 64 [0, 1] Geometric Representations Feature Description Range/Type bldg_poly_list List of building polygons Shapely polygons blk_utm_geometry Block boundary polygon Shapely polygon medaxis_geometry Block medial axis Shapely line The geometric data is extracted from OpenStreetMap and Microsoft Building Footprints. Medial axes are computed using a skeletonization algorithm. Socioeconomic and environmental features are taken from the U.S. Climate and Economic Justice Screening Tool, joined to blocks by spatially linking each block polygon to the census tract in which it resides.