UCID Benchmark Dataset v1.0: Temporal-Spatial Urban Context Assessment for 22 Cities (50,000 Instances)

This dataset contains 50,000 Urban Context Identifier (UCID) instances across 22 cities spanning Azerbaijan, the Nordic region, the Baltic states, and three benchmark cities (Istanbul, Portland, Sydney). Each UCID binds geographic coordinates (3-decimal precision), H3 hexagonal spatial indexing (resolutions 8-11), ISO 8601 temporal encoding (24-week window with 12 samples per location), and six urban quality contexts: 15-minute accessibility (15MIN), transit quality (TRANSIT), climate resilience (CLIMATE), urban vitality (VITALITY), equity analysis (EQUITY), and walkability (WALK). The dataset was generated using public data sources including OpenStreetMap (network topology and points of interest, snapshots from 2025-01 to 2025-03), GTFS transit schedules (6-month coverage 2024-12 to 2025-06), and census demographic layers. Context scores (0-100) are accompanied by letter grades (A+ to F), calibrated confidence estimates (0.00-1.00), and provenance flags for data lineage tracking. This benchmark enables reproducible research in urban analytics, supports transfer learning experiments across diverse urban morphologies, and provides standardized identifiers for longitudinal urban monitoring and cross-city comparisons. File: ucid_benchmark_v1.csv (50,000 rows × 14 columns)Columns: ucid, city, country, region, lat, lon, h3_res, h3_index, timestamp, context, raw_score, grade, confidence, flags Regional Distribution:- Azerbaijan (3 cities): 7,000 instances- Nordic region (10 cities): 23,300 instances - Baltic states (6 cities): 10,700 instances- Benchmark set (3 cities): 9,000 instances Context Distribution:- 15MIN: 10,500 instances- TRANSIT: 9,800 instances- CLIMATE: 7,200 instances- VITALITY: 8,900 instances- EQUITY: 6,100 instances- WALK: 7,500 instances For full methodology, provenance details, and evaluation protocols, see the accompanying arXiv preprint and GitHub repository.