Like any other discipline or science, urban planning is dedicated to the empirical study of planning and organizing cities and territories. It faces not only the challenges of designing agile mobility, planning accessible residential fabric, allocating land for economic activities, and efficiently managing environmental and energy impacts, but also the ultimate goal of making the resulting cities as livable as possible.
Named after Arturo Soria, Arturo is an algorithm trained by citizens and designed to determine which urban variables make our cities more livable. Much like Arturo Soria himself—whose commitment to livability was evident in projects like the Ciudad Lineal—the algorithm aims to turn livability into a measurable parameter, into objective data. To achieve this, it relies on the input of thousands of citizens to build an accurate picture of what makes a city livable.
The project asks citizens to evaluate side-by-side pairs of photographs of various city streets, aiming to generate a database of the highest- and lowest-rated spaces. Each image is backed by 50 intrinsic urban parameters, including built density, land use, building geometry, age, and construction quality. Once this set of images—representing just 10% of Madrid's urban fabric—is sorted, the project analyzes the shared characteristics of those with similar ratings.
This pattern-matching is executed by a machine learning algorithm (Gradient Boost), which identifies these trends and determines which of the underlying urban data points are most influential in classifying the photographs. Once the initial study is complete, the methodology can be scaled to cover the neighborhoods omitted from this first phase.
While similar initiatives exist—such as César Hidalgo's StreetScore or Daniele Quercia's Urbanopticon—this experiment offers, for the first time, an interpretation of the results based on the specific urban variables associated with each image. Ultimately, this approach recognizes that the visual experience of space is the most critical factor in shaping the urban environment.
This yields a more precise evaluation of the most decisive factors in urban planning, providing data-driven insights that can directly influence current planning tools. The entire study is freely downloadable, allowing anyone to use the data and further our collective understanding of cities through citizen-sourced information.
This article was written by Borja Fernández. The translation is powered by AI.
