Observing people behavior in public spaces.

When designers try to improve or make any changes in public space, they, firs of all, try to analyze people behavior and their interaction with that space. Today, there are numerous ways to do so. The first one and the most obvious way is going to the place and just spend time on observing people activities, movements, interactions with one another and space. For example, Jane Jacobs in her book “The Death and Life of Great American Cities” accurately described behavior pattern of the Greenwich Village neighbors, their habit and representativeness in space. However, the method that worked in 1960th is not useful now. Our pace of life doesn’t allow us to spend weeks and month on observing and manually fixating any changes in space.

RGB cameras for tracking people movements.

Fortunately, RGB and thermal cameras can help us to proceed faster and more effective. That small exercise aims to test how useful stable camera could be in tracking people movements. The public space, that I chose for that exercise is a square in front of market in La Barceloneta. That square is empty and, many economic activities surround it. There are supermarkets, restaurants, bakeries etc. Normally, people just go through without spending time on the square.  The 3 minute video recording happened on Sunday 12 pm, December. The camera overview catches only half of the square. The video was proceeded with YOLO object detection algorithm. Algorithm checked every 30th frame of the video that is equal to 1 frame every second.



Objects and their coordinates detected from algorithm appeared on the map. The following heatmaps depict the most popular passes and people concentration in each specific location.

concentration of people on a grid (3*3 meters)

heatmap of people movements

 Pros and cons.

The disadvantage of that method is that in order to get accurate result we should install our camera higher from ground, for example, on the rooftop of a building. Moreover, RGB cameras recordings arises question about privacy. The thermal cameras are more reliable in that sense as they don’t identify the person itself but the temperature difference. In addition, thermal cameras perform well during night when RGB are useless. The next step of that exercise will be attempting to identify the age group of the detected person and try to track each.


is a project of IAAC, Institute for Advanced Architecture of Catalonia
developed at Big Data Analytics e.g. Master in City & Technology,
in 2020/2021 by:
Students: Linara Salikhova
Faculty: Diego Pajarito