Last month we used on open source Computer Vision algorithms to count the number of pedestrians, bicyclists and vehicles entering and exiting a strategic intersection in the heart of Amsterdam. Tracking these objects reveals beautiful patterns - or ‘virtual graffiti’ - offering urban designers new insights into actual behaviors in our public space.
The study was conducted in the first two weeks of October 2019. See the interactive dashboard below to explore more details of the data, for example the peak hours, start and ending points, traffic distribution and more. The data was collected on a solar powered portable camera, and the open source code was improved to include bicyclists.
For a link to the Github repository go to: https://github.com/maartensukel/video-to-traffic-flow-information
Special thanks to Maerten Sukel - Lead AI developer @ City of Amsterdam & Aurore Paligot - Data analytics and visualization.