The Urban Big Data Centre (UBDC) is working closely with Glasgow City Council (GCC) to develop innovative methods for monitoring activity on Glasgow’s streets using excess capacity on the city’s extensive CCTV network.

Using machine learning techniques, we have been applying object detection models to identify pedestrians, cyclists, and seven types of vehicles.

Data from the project is freely available through a UBDC API and GCC’s open data platform.


In developing this project, we had two objectives:

  • To test whether it was possible to develop methods to measure activity on streets across Glasgow within a CCTV environment while ensuring that sensitive image data did not leave the secure systems.
  • To gather data for use in the ongoing evaluation of Glasgow’s Avenues project - a public realm project designed to reduce traffic, encourage walking and cycling, and increase economic activity in the city centre.


The project has developed in several ways that have proved valuable to GCC by providing data to which they otherwise wouldn’t have had access. The team at UBDC has benefitted from the opportunity to continue developing these methods. 

We now routinely collect data from 40 cameras, capturing and processing images every 30 minutes. As well as providing ongoing data to GCC, the project also enabled the City Centre Task Force, which was launched to tackle the effects of the Covid crisis, to monitor and compare activity levels during and after the COVID-19 pandemic. 

We have also been able to provide an analysis of activity in Kelvingrove Park, including information about peak use times, which has helped GCC to manage access and plan the response to large gatherings.

As well as providing information on activity levels, the data can be used to estimate where there have been high numbers of pedestrians which the council hopes to feed into optimising routing for street cleansing. 

In addition to facilitating several projects and routine council planning, we have also received requests to provide pedestrian counts for North Glasgow and have offered to use the model on footage from external cameras. We also provided data to the Office of National Statistics (ONS) through our API to add to data collated to inform government planning during the pandemic.

As well as applying and optimising open object detection models, we have developed video analytics, which allows us to carry out bidirectional counts on CCTV videos and trace the paths of individuals through the streetscape. Both methods give us different tools to observe changes in how people use the street once new infrastructure is put in place. The most recent work has involved working with GCC’s Neighbourhoods, Regeneration and Sustainability team to develop methods for estimating parking occupancy on selected streets. This project will examine parking occupancy, collecting data every 30 minutes. The pilot will run over 2-3 weeks, and we will manually check the accuracy of the predictions from a sample of the records.

This project reinforces the benefits of identifying opportunities for collaboration between academic researchers and local authorities on projects that utilise urban data. From UBDC’s perspective, the project has academic value and provides a data resource for our data service. For GCC, the project provides them with valuable information for planning and service delivery. This mutually beneficial relationship is vital for the success and continued impact of the project.

Kimberley Hose, Head of Business Intelligence at Glasgow City Council, welcomes ongoing collaboration with UBDC. She said: “Local authorities deliver a myriad of services and, because of cuts in recent years and budget constraints, we have to focus very carefully on where best to place our resources. Using big data effectively and efficiently, in collaboration with Data Scientists, is absolutely critical to this process in support of the council and city transformation."

Data from the project is freely available through a UBDC API and GCC’s open data platform.