Proposals answered within 2 business days
All case studies
Smart cityGDPRIoT

Smart Cities: traffic, energy and safety in one system

Client: Mid-sized city (anonymised data)Industry: Public sector18 February 2026
Results
travel time on main corridors
-9%
energy use by street lighting
-15%
images of people stored outside the camera
0

The challenge

The city has about 120,000 residents. Traffic signals, street lighting and CCTV ran on three separate systems, managed by different units. None of them could see the others' data, so traffic jams, lamp failures and incidents were analysed separately and with a delay.

GDPR was an extra constraint. The public safety department wanted to use cameras to count traffic and detect events, but the data protection officer rightly asked what exactly is processed and where it goes. Without a good answer, the project would not have moved an inch.

The approach

We designed the system so that personal data never leaves the edge devices. Image analysis happens in the camera or in the cabinet at the junction. Only aggregated numbers reach the platform: how many vehicles, of what type, at what speed, how many people at the crossing. The image is deleted after processing.

Before the first deployment, we ran a data protection impact assessment (DPIA) with the city and agreed the scope of processing with the officer. On the same platform we combined data from traffic sensors, lamps and the power grid, so that one team sees the whole picture.

The rollout

The pilot covered two main corridors and 23 junctions. The algorithm adjusts signal cycle lengths to current traffic volume, and priority for city buses is granted automatically. At the same time, street lighting at more than 3,000 light points got adaptive dimming based on traffic and time of day.

Operators work in a single dashboard: traffic, energy, failures and incidents. Data access is split by role, and every query leaves a trace in the audit log. The whole project took 11 months, 4 of them for the pilot.

The results

On the pilot corridors, average peak-hour travel time fell by 9%. That sits within the range published studies report for adaptive signal control. Lighting energy use dropped by 15% against the previous season, even after the earlier switch to LED.

The safety department now sees road incidents almost live, without storing anyone's likeness. The city plans to extend the system to more corridors in 2027.

Technologies

  • NVIDIA Jetson
  • Apache Kafka
  • TimescaleDB
  • Kubernetes
  • Python
  • OpenCV
  • Grafana
  • Keycloak