clean_trajectory
Use this skill when the user wants to clean raw GPS, check-in, signaling, taxi, or user mobility trajectory files by validating coordinates and timestamps, removing duplicates, filtering abnormal spee
SKILL.mdskills/custom/spatiotemporal_trajectory/clean_trajectory/SKILL.md
GitHubclean_trajectory
Use this Skill for the first step of the trajectory pipeline: raw trajectory points -> cleaned trajectory points.
Command
cd /mnt/skills/custom/clean_trajectory
python3 scripts/clean_trajectory.py --input /path/to/raw.csv --output-dir /path/to/output
Optional flags:
--user-col user_id --trajectory-col trajectory_id --time-col timestamp --lat-col lat --lon-col lon
--max-speed-kmh 200 --geohash-precision 6
Outputs
cleaned_points.jsonlsummary.json
Algorithm
- Coordinate validation
- Timestamp normalization
- Duplicate removal
- Haversine distance
- Speed-threshold filtering
- Geohash encoding
Next Skills
Use cleaned_points.jsonl with:
$map_spatial_grid$search_spatiotemporal$analyze_region_heat$mine_trajectory_patterns