Repository navigation
Traffic Flow Analysis Version 1.0.0
This is the first release for Traffic Flow Analysis.
More detail about the app can be found in
Data Sharing API
* `/api/v2/count` - Endpoint for vehicle count prediction
* `/api/v2/count/<vehicle_type>` - Endpoint for vehicle count prediction, support 3 types of vehicle: `car`, `motorcycle`, and `heavyvehicle`
* `/api/v2/flow` - Endpoint for traffic flow speed prediction
* `/api/v2/desc/flow` - Endpoint to get average flow data over the time from database, aggregated into 30 interval time
* `/api/v2/desc/count` - Endpoint to get number of average vehicle count each type over the time collected in database, aggregate into average
* `/api/v2/desc/weather` - Endpoint to get number of weather condition collected each class. Weather condition get interpret into 4 classes: `Clear`, `Cloudy`, `Rain`, `Low Visibility`
Frontend Web Application
* Data visualization - Visualize historical and predicted data using charts
* API Exploration - Explore and interact with the prediction API endpoints from within the web app.
Prediction Model
* Prediction model based on Recurrent Neural Network (RNN).
* It is exported from PyTorch and run using the LibTorch library in Rust.
* The model is pre-trained and doesn't update automatically, model retrain is require to make the model change it prediction to align with newly added data.