Urban Heat, Urban Sales: A Neighborhood-Scale Dataset of Retail Heat Sensitivity in Seoul

This dataset provides a city-scale, daily-resolution view of how retail commercial activity in Seoul responds to extreme heat. It covers all 422 administrative neighborhoods (dong) of Seoul for the year 2024 (366 days), and links daily weather, urban structural features, and a Retail Heat Sensitivity Index (RHSI) that measures the change in retail transaction volume between extreme-heat days and mild days for each neighborhood. The dataset was developed to support operational monitoring of heat-related economic vulnerability by identifying which commercial districts are most sensitive to heat. It is intended for research and planning use in urban climate, retail economics, and geospatial analysis. The dataset was constructed through a data partnership with the Seoul Metropolitan Government and the Seoul AI Foundation, which provided access to proprietary daily credit-card transaction records. All released files contain aggregated neighborhood-level variables only; they do not include individual-level, merchant-level, or raw transaction records. The dataset consists of five files: RHSI.csv — The Retail Heat Sensitivity Index (RHSI) and supporting day counts for each of the 422 dong. RHSI is the log-ratio of mean daily retail transaction volume on hot days (apparent temperature ≥ 33°C) to mild days (18–26°C), capturing how heat suppresses or shifts local retail activity. Urban_Features.csv — Neighborhood-level urban structural features per dong, spanning retail/sales composition, population dynamics, demographics, land use, and accessibility. These are the explanatory variables used to characterize where retail heat sensitivity concentrates. Daily_Weather.csv — Daily weather records for each dong over the 366-day period, including maximum air temperature, apparent temperature, humidity, precipitation, and day-type flags (mild day, hot day, holiday). Administrative_Dong_Geometry.geojson — Polygon boundaries of the 422 administrative dong (EPSG:4326), used for spatial joins and mapping. Data_schema.csv — A data dictionary documenting every column across all files: name, file, data type, unit, missing-value handling, definition, construction formula, source, and interpretation. All tabular files share dong_code as the common spatial key, allowing them to be joined into a single neighborhood-level dataset.