Chicago_COMPASS Community area summaries for Chicago and related scripts. Specifics below: Data Each CSV file provides summaries for 77 community areas in Chicago from WRF simulations, satellites, and socioeconomic surveys. The spatial polygons for the community areas and the socioeconomic data were accessed through the Chicago Data Portal: https://data.cityofchicago.org/ The WRF code is open source and can be found at: https://github.com/wrf-model/WRF Chicago_control, Chicago_no_urb, and Chicago_no_lake have the maximum and minimum average variables of interest for the control, no urban, and no lake simulations. These are for the BEM/BEP runs with the YSU boundary layer scheme and are used for the main results of the paper. The WRF_BEM_MYJ files are for the BEM/BEP control runs with the MYJ boundary layer scheme. The WRF_Noah files are the control runs using just the Unified Noah land surface model (no urban canopy). The WRF_nested file is for a control run using 3-way nested domains, with the inner domain over Chicago at 1.333 km using BEM/BEP and the YSU boundary layer scheme. Chicago_perc_control, Chicago_perc_no_urb, and Chicago_perc_no_lake have the 95th and 98th percentiles of hourly variables of interest for the control, no urban, and no lake simulations. Chicago_MODIStime_control has the daytime and nighttime variables of interest (corresponding to MODIS Aqua overpass) for the control simulations. en01, en02, en03, and so on represent the ensembles for each model configuration. Chicago_geo_socioeconomic includes the socioeconomic variables (median income per capita and Hardship Index), spatial metrics (area and distance from the coast), and satellite-derived estimates (daytime and nighttime land surface temperature (LST), and normalized different vegetation index (NDVI). Scripts WRF_to_tabular.R converts the WRF simulations into tabular data to be injested into Google Earth Engine. Rasterize.js converts the tabular WRF results into a raster with separate bands for each variable on Google Earth Engine. Summarize.js processeses satellite observations and summarizes the satellite and WRF outputs into regions of interest on Google Earth Engine.