County-Level US Multi-Modal Spatiotemporal Urban Growth & Travel Behavior Dataset (2012\u20132023)

This dataset comprises approximately 7,100 satellite images paired with corresponding demographic and travel behavior data spanning 2012-2023 (excluding 2020) across United States counties. The satellite imagery consists of 256\u00d7256 pixel Landsat 8 Collection 2 Level 2 surface reflectance composites covering 10 km\u00b2 areas around county centroids, processed to create cloud-free annual median representations. Demographic data includes 25 key variables from the U.S. Census Bureau's American Community Survey (ACS) 1-year estimates, encompassing population statistics, age distributions, racial composition, and educational attainment levels. Travel behavior metrics capture transportation modes, commute patterns, vehicle availability, and temporal travel characteristics for counties with populations exceeding 65,000. This multimodal spatiotemporal dataset enables research at the intersection of remote sensing, urban planning, and transportation analysis, providing a unique resource for studying the co-evolution of built environments, demographic patterns, and mobility behaviors over an 11-year period. The dataset supports applications in predictive modeling, urban development forecasting, transportation planning, and socioeconomic analysis using machine learning and computer vision techniques.