traffic-anomaly-detection
Link prediction and anomaly detection for traffic networks.
Plan
- Use STGAN, repository, my fork.
Instructions
- Install mamba/miniforge3.
curl -L -O "https://github.com/conda-forge/miniforge/releases/latest/download/Miniforge3-$(uname)-$(uname -m).sh"
bash Miniforge3-$(uname)-$(uname -m).sh
- Install dependencies.
mamba env create -f env.yaml
- Activate environment.
conda activate env
Traffic data
-
Dropbox links: confidential.
-
Notebook:
traffic_data.ipynb
.
Data
-
The X coordinate is the time, e.g.
0458
is 4:58am, and2038
is 8:38pm. -
The Y coordinate is the density.
Datasets
-
dataset1
: Data for every Monday, for 18 days. -
dataset2
: Data for consecutive days, including weekends, for 6 months, February through July. -
dataset3
: Data for consecutive days, including weekends, for 6 months, April through November.
Weather data
-
Dropbox links: confidential.
-
Notebook:
weather_data.ipynb
.
Ground Truth Data
-
Social events
-
Traffic accidents
-
Road closures