MUTra‑CDMX: Multisource Urban Traffic for CDMX

The primary dataset file, MUTra_CDMX.csv, comprises a total of 691,200 records organized in a rigorous spatiotemporal tabular format. Each observation denotes the state of a discrete road segment at a specific timestamp, thereby constituting a fully continuous time series devoid of missing values. This methodological architecture renders the dataset highly suitable for time-series analysis and the training of machine-learning algorithms for travel-time prediction. The dataset encompasses 20 consecutive road segments situated along an urban corridor, spanning an uninterrupted period of 120 days, from 1 November 2024 (00:00:00) to 28 February 2025 (23:55:00). Data was collected at a five-minute sampling frequency, yielding 288 observations per segment per day, ensuring comprehensive 24-hour coverage throughout the four-month study duration. The repository incorporates 20 variables, classified as follows: Two spatiotemporal identifiers: segment, datetime One target variable for predictive modeling: traveltime Seventeen predictor variables detailing traffic conditions, geometric attributes, urban environment characteristics, derived indicators, and meteorological factors. The geometric and contextual variables (distance, frc, signals, schools, hospitals, sinuosity, gradient, capacity) represent fixed attributes of each road segment, enabling structural comparisons across the corridor. The traffic-related variables (speed, freespeed, freetime, ivr, congestion, road_closure) characterize the operational state of the corridor at any given timestamp. Furthermore, the meteorological variables (temperature, rain, wind) supply pertinent environmental data essential for modeling traffic dynamics. The supplementary data dictionary provides comprehensive English-language descriptions, data types, measurement levels, and corresponding units, thereby facilitating the accurate interpretation and subsequent reuse of the repository.