GRACTRANET

This dataset contains structured traffic trajectory data extracted from Waze API JSON responses, designed for use in travel time estimation and traffic prediction tasks. Each record corresponds to a route query and includes multiple route alternatives with detailed metadata such as real-time, historical, and free-flow travel times. The dataset captures fine-grained spatial and temporal information, including latitude-longitude coordinates of road segments, segment distances, travel times, and navigation attributes like turn penalties and road classifications.To facilitate machine learning applications, the raw JSON data are parsed and converted into tabular and serialized formats consistent with deep learning trajectory modeling frameworks such as DeepTTE. Each trajectory includes spatial sequences (lats, lngs, dist_gap), temporal features (time_gap, weekID, timeID), and the ground-truth total travel time (time). This structure enables the modeling of dynamic traffic patterns across different times of day and days of the week. The dataset supports research in route optimization, ETA (Estimated Time of Arrival) prediction, and intelligent transportation systems.