Routing Server

A traffic-aware routing engine for Brno, built as part of the bachelor's thesis Routing Algorithm for Traffic Planning in Brno (BUT FIT, 2024/2025). It combines OpenStreetMap road network data with historical Waze traffic-jam records to compute routes whose cost reflects how congested the roads actually were during a given time window, rather than just their free-flow speed.

Given a source and destination coordinate plus a date range, the server returns the fastest path between them, its estimated travel time both with and without historical traffic, and a street-by-street breakdown colored by congestion severity — ready to be rendered on a map.

How it works

• Map data (osm.py, graph.py) — road geometry for an area (e.g. "Brno") is pulled from the Overpass API and converted into a networkx directed graph, with edges weighted by per-road-type speed limits (motorway, residential, service, etc.).
• Traffic data (traffic.py) — historical Waze jam records are loaded from PostgreSQL, spatially matched against graph edges, and used to compute a traffic-adjusted traversal time and severity (0–2) for each road segment, based on how often and how badly it was jammed within the requested date range.
• Routing (routing.py) — shortest paths are found with A*, using an ALT heuristic (A*, Landmarks, Triangle inequality) precomputed from a set of landmark nodes, which is much faster than plain straight-line A* on a real road network.
• API (main.py) — a FastAPI service that builds the graph and traffic overlays on startup, keeps a cache of traffic-adjusted graphs per date range (refreshed daily and pre-warmed for the last 7 days), and exposes an endpoint to compute routes on demand.

The architecture is geography-agnostic — swapping the AREA constant is enough to route in a different city, as long as matching Waze traffic data is available.

API

POST /find_route_by_coord

Finds the fastest route between two coordinates for a given date range.

Request body:

{
  "src_coord": [16.6068, 49.1951],
  "dst_coord": [16.5765, 49.2108],
  "from_time": "2025-05-01",
  "to_time": "2025-05-07",
  "use_traffic": true
}

Response:

{
  "streets_coord": [
    {"street_name": "Údolní", "path": [[lat, lon], [lat, lon]], "severity": "0", "color": "green"},
    {"street_name": "Kounicova", "path": [[lat, lon], [lat, lon]], "severity": "1", "color": "orange"},
    {"street_name": "Veveří", "path": [[lat, lon], [lat, lon]], "severity": "2", "color": "red"}
  ],
  "route": [[lat, lon], ...],
  "src_street": "Údolní",
  "dst_street": "Veveří",
  "length": 1234.5,
  "time_with_traffic": 210.3,
  "time_without_traffic": 180.0
}

Running the server

Prerequisites

• Database credentials in .env file
  • DB_HOST
  • DB_PORT
  • DB_USER
  • DB_PASSWORD
  • DB_NAME
• Python 3.12+ for local setup
• Docker and Docker Compose for Docker setup

Local setup

pip install -r requirements.txt
chmod +x run.sh
./run.sh

Docker setup

docker compose up -d