Overview

The Glasgow CCTV Object Detection platform provides counts of persons, cyclists and different types of vehicles for more than 50 locations across the city. This counting process is achieved by using the Glasgow City Council network of CCTV cameras. Every 30 minutes a snapshot is taken at the same location, and the resulting image is processed by a detection model to localise the objects of interest on it. Afterwards, the localised objects are counted and published in the API. See example below:

Logging new data at regular intervals each day, this project aims to improve public understanding of mobility and traffic patterns across Glasgow.

This platform provides data from two object detection models:

1. An off-the-shelf object detection model, provides counts of persons, cars, buses, trucks, bicycles and motorcycles. This data has been recorded since October 2019.
2. An in-house developed Yolo model provides counts of buses, cars, cyclists, crowds, lorries, motorcycles, pedestrians, taxis and vans. This data has been recorded since January 2023.

Getting started

The data in the API can be accessed in two ways:

1. API request, which allows a higher level of filtering (by camera and day of the year). For detailed explanation and endpoints available, please check our API reference page.
2. Downloading a csv file with the data for all cameras and for all days. To accomplish this, press the download button available on the main page:

You can also visualize the data with our dashboard and perform aggregations by period of time and using different aggregation methods. Furthermore, aggregations can be filtered by classes of objects and by camera.

---

Note:

Download and dashboard functionalities are only available for the Yolo model.

---

How-to Guides

Python Example

Please check this GitHub repo for examples in Python programming language.

R Example

Please check this GitHub repo for examples in R programming language.

Want to know more?

Please check our technical notes at: