Background
The Prague 18 (Letňany) municipal district commissioned continuous traffic monitoring across three of its busiest junctions from November 2022 to March 2023. The district had a specific question to answer: whether real traffic loads were higher than the figures quoted alongside a planned development. It also wanted a reliable dataset for future transport planning.
Letňany carries steady transit traffic because the approach routes from the D8 and D10 motorways (formerly R10) have limited capacity. Parking is the second pressure point, with commuters driving into Prague filling local capacity. For a district developing this quickly, resident safety and workable transport service both depend on knowing what the traffic actually does.
The Challenge
The district needed evidence that would hold up in a planning discussion, and a short count campaign could not provide it:
- Contested figures: the declared traffic loads needed independent verification, not another estimate
- No continuity: repeated manual counts give a few days of data per campaign and nothing in between, so trends stay invisible
- Event distortion: major events in the district move volumes sharply, and a short count either misses that entirely or gets skewed by it
- Detail: direction-level flows, vehicle classes, speeds and turning movements at a roundabout are hard to capture manually at any useful scale
- Privacy: continuous camera monitoring in a residential district cannot depend on storing video footage
The Solution
Three locations were selected for monitoring:
- Toužimská × Beranových
- Tupolevova × Beranových
- Tupolevova × Veselská (roundabout)
The three camera sites in the GoodVision platform
Each site got one 4 MP IP camera with an internet connection. KH servis integrated the camera sites. GoodVision supplied the traffic intelligence: a local edge unit, AI processing, analytics configuration, the cloud platform, support, and regular reporting.
Camera streams were analyzed locally in real time. Only anonymized data traveled to the GoodVision cloud over a secure connection, covering vehicle type, direction, speed, time and movement trajectory. No video was stored. Prague 18 had direct access to live results and received regular summary outputs alongside them.
Trajectory analysis at the Tupolevova × Veselská roundabout
What GoodVision Live Traffic Delivered Automatically
- Traffic volumes by direction and time of day
- Vehicle classification using our 8-class AI Engine
- Turning movements, speeds and trajectories
- Daily and monthly trends
- Comparison of normal traffic against major event periods
- Dashboards, reports and data exports
One camera covered several directions and traffic classes at once. Analytics profiles were edited in software, with no return visit to the camera site.
Tupolevova × Beranových, several directions and traffic classes from one camera
Results & Impact
Across five months, the two main entry and exit profiles recorded 3.6 million vehicle passages between them:
|
Metric |
Result |
|
Monitoring period |
5 months continuous (November 2022 to March 2023) |
|
Hardware |
3 standard 4 MP IP cameras, one per junction |
|
Vehicle passages, Tupolevova |
2.50 million |
|
Vehicle passages, Toužimská |
1.11 million |
|
Motocykl 2023 event period |
83,358 passages, 13.7% above the comparable Gaudeamus period (73,336) |
|
Motorcycles, same comparison |
43 to 347, more than eight times higher |
|
Video stored |
None |
Monthly traffic on Toužimská, November 2022 to March 2023
Monthly traffic on Tupolevova, November 2022 to March 2023
What the Event Comparison Showed
The Motocykl 2023 analysis is the clearest example of what continuous measurement adds. Total traffic on the Tupolevova profile rose 13.7% against the Gaudeamus reference period, which on its own is a moderate change. The class breakdown tells a different story: motorcycles went from 43 to 347, more than eight times higher. A count campaign scheduled on a normal week would have reported neither figure, and one scheduled during the event would have reported the peak as the norm.
Gaudeamus 2023 (73,336 passages) against Motocykl 2023 (83,358) on the Tupolevova profile
Because the system ran without interruption, the district ended up with a consistent picture of real traffic that covers weekdays, weekends, long-term trends and the effect of major events in the same dataset.
Inspiration for Other Cities
The project shows that standard camera infrastructure works as an accurate, uninterrupted source of traffic data. Compared with repeated manual counting, it gives higher consistency, far more detail, and results available as they happen, without storing any video.
The same principle applies to pedestrians, cyclists and other forms of micromobility. Analyzing trajectories, speeds and interactions between road users helps reveal conflict points and risky situations before they turn into collisions.
That gives cities and municipalities objective evidence for traffic and parking planning, for developing walking and cycling infrastructure, for assessing the safety of vulnerable road users, and for measuring whether the measures they have already built are working.
Project Roles
- Prague 18 – Letňany municipal district: client and data user
- KH servis: integration of the camera sites
- GoodVision: supplier of AI traffic intelligence and the analytics platform
Book a demo at goodvisionlive.com/request-demo/ and see what your existing cameras can measure with GoodVision Live Traffic.
About GoodVision
GoodVision is an AI-powered traffic analytics platform serving 200+ transportation authorities globally across four continents. Our technology transforms video footage into actionable traffic insights, helping cities and consultancies make data-driven decisions about urban mobility and infrastructure development.