Every call you place is creating a map of your city, and it could potentially help solve our traffic problems.

      You don't need to explicitly share your location for your city to determine where you are. Whenever you make a call or send a message, you connect to a nearby network antenna. When this activity is multiplied by millions of people engaging in the same actions daily, the result is not just data — it creates a dynamic and accurate representation of a city's functioning. Researchers at the University of Córdoba have successfully tapped into this with a new tool designed to analyze these patterns.

      The tool that observes without actually observing

      MAPLID (Multi-label Approach for Place Identification) does not track specific individuals. Instead, it examines patterns — aggregated and anonymized signals that reflect how locations function over time. It can indicate when an area transitions from residential to commercial use, when traffic to industrial zones peaks, or how a significant event can subtly alter the flow of an entire neighborhood.

      What sets it apart is its ability to capture the multiple facets of a location. For instance, a university campus is not solely a workplace; depending on the time of day, it can also serve as a home, a social hub, or a transit corridor. Unlike most mapping tools that tend to assign a single label, MAPLID encompasses all these dimensions simultaneously.

      Understanding the research methodology

      The model was developed as part of doctoral research by Manuel Mendoza Hurtado, alongside colleagues Juan A. Romero del Castillo and Domingo Ortiz Boyer from the Department of Computer Science and Artificial Intelligence.

      Rather than relying on raw location data, the system builds its understanding in layers. It begins with geolocated metadata from calls and messages — not the content itself, but the connection points that are triggered when devices connect to network antennas. It then analyzes how these signals vary over days and weeks, helping distinguish routine behaviors from isolated movements. This behavioral data is then aligned with OpenStreetMap, an open-source geographic database, providing real-world context such as street types, landmarks, and building classifications, thereby transforming abstract signal patterns into a more practical format for urban analysis.

      The outcome of this approach is akin to a time-lapse. The same street block can tell entirely different stories at various times — for example, 7 AM looks vastly different from 7 PM. To evaluate the model, the team applied it to Milan and Trento, two Italian cities that differ significantly in size and layout, making them ideal for comparison. Due to privacy regulations, Spanish mobile data was unavailable, so the researchers utilized a dataset provided by Telecom Italia for academic research. Despite layering millions of daily data points onto urban maps, the model performed consistently across both cities, indicating its applicability across diverse urban environments.

      The study has been published in the International Journal of Geographical Information Science.

      So, who is actually monitoring this?

      Currently, no one is officially utilizing it. The researchers’ next step is to present the tool directly to local governments and city planners. The potential applications are quite evident — adjusting bus schedules according to actual movement patterns, enhancing traffic flow in congested areas, and dispatching cleaning crews to locations that truly require their attention, rather than relying on outdated assumptions.

      The intriguing aspect is that cities have always generated this type of information; it has never been absent. What has been lacking is a method to interpret it in a meaningful and practical manner. This tool could very well be the advancement needed to bridge that gap.