This project implements a real-time dynamic pricing system for smart parking management using data streaming with Pathway and visualizations via Bokeh and Panel. It dynamically adjusts parking prices based on real-time data such as occupancy, traffic, queue length, and special events to maximize efficiency and revenue.
We simulate data streaming from historical records and plot real-time pricing graphs for each parking lot using Bokeh-powered interactive dashboards.
| Layer | Tools / Libraries |
|---|---|
| Language | Python 3 |
| Stream Engine | Pathway |
| Visualization | Bokeh, Panel |
| Data Format | CSV |
| Platform | Google Colab |
The following diagram illustrates the real-time dynamic pricing system architecture :
🔗 Click on the image to view it in full resolution.
- Data Ingestion We read a CSV file (dataset.csv) simulating historical parking data.
Pathway reads it in streaming mode, processing records at fixed intervals (autocommit_duration_ms=500).
- Real-Time Processing (Model 2) A custom UDF (compute_dynamic_price) computes the parking price based on:
Occupancy rate
Queue length
Traffic conditions
Vehicle type
Special event flag
This price is continuously calculated and updated as new rows arrive.
- Real-Time Visualization We use Pathway's .plot() method to bind real-time data to a custom Bokeh chart.
Panel serves this plot as an interactive live dashboard.