Multiview Vehicle Mobility in Smart Villages Dataset: Graph Trajectories, Tabular Context, and Temporal Series

Abstract A comprehensive multiview dataset for vehicle mobility analysis and repeater prediction. The dataset includes three complementary views: Graph View. Vehicle trajectories represented as sequences of nodes (locations) with timestamps and directions. Tabular View. Contextual features for each visit, including total distance, duration, overnight stays, and holiday indicators. Temporal View. Daily aggregated traffic flow time-series data spanning from 2022 to 2024. Variables: The dataset includes three complementary views: Graph View. Vehicle trajectories represented as sequences of nodes (locations) with timestamps and directions. num_plate: Anonymized unique identifier for each vehicle. route: List of camera nodes visited during the trip (e.g.,['NodeA', 'NodeB']). times: List of time intervals (in minutes) associated with the traversal between nodes. directions: List of binary indicators representing the direction of travel between nodes. entry_date/exit_date: Timestamp of the first and last detection of the visit. repeater: Binary target variable indicating whether the vehicle returned to the system in the future (True/False). Tabular View. Contextual features for each visit, including total distance, duration, overnight stays, and holiday indicators. total_distance: Estimated total distance traveled by the vehicle during the visit. visit_time: Total duration of the visit in hours. nights: Number of nights the vehicle stayed (inferred from entry/exit dates). entry_day_of_week/exit_day_of_week: Day of the week for the entry and exit events. entry_is_weekend/exit_is_weekend: Boolean indicating if the entry/exit occurred on a weekend. entry_is_holiday/exit_is_holiday: Boolean indicating if the entry/exit occurred on a national holiday. entry_is_high_season: Boolean indicating if the visit occurred during the peak tourism season (e.g., August). entry_hour/exit_hour: Hour of the day (0-23) for entry and exit. Temporal View. Daily aggregated traffic flow time-series data spanning from 2022 to 2024. day: Date of the observation (YYYY-MM-DD). count: Total number of unique vehicles detected in the system on that specific day.