ewha-cong_dynamicP

A Python module for calculating and analyzing three types of congestion—Road, PM, and Pedestrian congestion—using trajectory data collected from CCTV systems. It includes safety-aware adjustments based on pedestrian-vehicle and pedestrian-pedestrian interactions.

How to use

'''python from ewha_congestions.three_congestions import processing_congestions

df = your data. data index must be DatetimeIndex as "dtct_dt". data must has point geometry as "geomatry", "snr_id", "distance", "speed", "acceleration", "traj_id", "direction", "mf_type", and "apr_code"

road_shp = your road data. road data must has Polygon geometry of roads as "geometry" and CCTV ID as "snr_id".

road_shp = road_shp.to_crs("EPSG:5179") for calculate road area

road_cong, ped_cong, pm_cong = processing_congestions(df, road_shp, 3600).call()

''' df: includes object tracking data with fields like snr_id, mf_type, apr_code, geometry, dtct_dt, speed, and acceleration.

road_df: includes road information with fields snr_id and geometry.

interval: time window (in seconds) over which congestion is aggregated and computed.

Output

Each output (road_cong, ped_cong, pm_cong) is a DataFrame with:

CCTV_ID: unique identifier for each camera

time: timestamp of the congestion record

congestion_level: integer from 0 to 4

0: Excellent (≤ 0.2)

1: Good (≤ 0.4)

2: Normal (≤ 0.6)

3: Crowded (≤ 0.8)

4: Very Crowded (> 0.8)

Congestion Types Type | Description Road Congestion | Combined area occupied by all moving objects divided by the road polygon area PM Congestion | PM area only, normalized by unoccupied space (excluding pedestrian & vehicle) Pedestrian Congestion | Road congestion adjusted using pedestrian-pedestrian and pedestrian-vehicle conflicts

Requirements

numpy pandas geopandas movingpandas tqdm shapely geopy

License

This project is licensed under the JiyoonLee License.

Contact

0197black@gmail.com