I built a live traffic intelligence dashboard for a…
0 reactions
I built a live traffic intelligence dashboard for a highway corridor using AI-powered camera counting
I set up a camera overlooking North Coast Highway (PCH) in Oceanside, California and built a system that uses YOLOv8 + ByteTrack to
count vehicles, pedestrians, and bicycles in real-time, 24/7.
The dashboard shows:
\- Live camera feed with 3-second refresh
\- Vehicle vs pedestrian counts with directional split (northbound/southbound)
\- 30-minute traffic flow chart
\- Day-of-week heatmap showing when the corridor is busiest
\- Best visibility windows ranked by hour
\- Live weather, tides, and sunset data
Stack: Go API, Python/YOLO vision worker on an RTX 4060, PostgreSQL, Cloudflare Tunnel. Everything runs on a home server.
Data collection started April 5, so the heatmap is still building up — patterns will get more interesting over the next few weeks.
Note: This counts aggregate movement only. No individuals are tracked, no faces or plates are captured or stored.
#education
source
I set up a camera overlooking North Coast Highway (PCH) in Oceanside, California and built a system that uses YOLOv8 + ByteTrack to
count vehicles, pedestrians, and bicycles in real-time, 24/7.
The dashboard shows:
\- Live camera feed with 3-second refresh
\- Vehicle vs pedestrian counts with directional split (northbound/southbound)
\- 30-minute traffic flow chart
\- Day-of-week heatmap showing when the corridor is busiest
\- Best visibility windows ranked by hour
\- Live weather, tides, and sunset data
Stack: Go API, Python/YOLO vision worker on an RTX 4060, PostgreSQL, Cloudflare Tunnel. Everything runs on a home server.
Data collection started April 5, so the heatmap is still building up — patterns will get more interesting over the next few weeks.
Note: This counts aggregate movement only. No individuals are tracked, no faces or plates are captured or stored.
#education
source
| 3,000 mlx | |
| 3 | |
| 0 |