How can citizens train an algorithm?

Arturo is a citizen science experiment designed to determine what characteristics make a city habitable by training an artificial intelligence algorithm.

The notion of habitability has been redefined throughout history by multiple city planners. In this project, we approached it from a collective perspective to building objective, recyclable knowledge from citizens’ individual perceptions.

Citizens are invited to an online participation platform where they choose between pairs of photos of different streets. In doing so, an artificial intelligence algorithm is trained through a big data model with urbanistic variables associated to the images selected by the citizens.

Through this process, we can identify the most habitable places and relevant urbanistic parameters and offer a view of gender, age and origin as well thanks to the segmentation of the data collected. As of today, more than 4,000 citizens have placed some 95,000 votes in the Madrid pilot study.

Related links



Fundación COTEC


In the news

Related expertises