What problem does it solve?
Provides a repeatable workflow to clean noisy GPS/trajectory data, detect and remove drift and redundant points, smooth and segment trips, identify stay and move segments, and prepare trajectories for map-matching and analysis so that urban mobility datasets become reliable for downstream analytics.
Core Features & Use Cases
- Drift and Redundancy Removal: Remove speed/distance/angle outliers and duplicate points to improve data quality.
- Smoothing & Segmentation: Apply Kalman smoothing and time-gap based segmentation to reconstruct continuous trips.
- Stay/Move Detection & Trajectory Densification: Identify stay points, densify or sparsify samples, and convert points to linestrings for spatial analysis.
- Map Matching & Length Calculation: Integrate with OSMnx road graphs for map-matching and compute trajectory lengths using Haversine or Euclidean methods.
- Use Case: Clean taxi GPS logs, detect passenger pick-up/drop-off stays, densify traces for routing, and map-match to road networks for travel behavior studies.
Quick Start
Use the transbigdata-traj skill to clean a taxi GPS CSV, detect stay points with a 30-minute threshold, and convert results into trajectory lines for visualization.