Airbnb and Urban Renewal in Lisbon (2015–2019): Spatial Interactions Revealed by Network-Enhanced Kernel Density Modelling

Cities are living maps, constantly being redrawn by the movements of people and the flow of money. For decades, urban planners have tried to understand how different forces shape a city, often by looking at where things happen. When a new business opens or a building is renovated, it leaves a mark on the landscape. In recent years, a powerful new force has appeared on these maps: short-term rental platforms like Airbnb. These digital marketplaces allow homeowners to rent out their apartments to travelers, creating a massive, invisible network of temporary housing that operates at a very fine scale, down to individual streets. At the same time, many European cities have been undergoing intense cycles of building renovation, as owners and investors work to restore historic structures. The question for researchers is not just whether these two things are happening, but how they relate to one another in physical space. Do they occupy the same neighborhoods? Do they follow the same streets? To answer this, scientists use tools that turn scattered points of data into smooth, continuous pictures of intensity, showing where activity is most concentrated. However, cities are not flat, open fields; they are complex webs of streets and intersections. A method that treats the city as a flat surface can sometimes blur the details, making it hard to see exactly which streets are carrying the most pressure.

A new study focused on Lisbon, Portugal, between 2015 and 2019, to untangle this relationship. During these years, the city saw a rapid surge in Airbnb listings alongside a heavy cycle of licensed building rehabilitation. The researchers wanted to see if these two phenomena were simply overlapping in the same general areas or if they were interacting in more specific, street-level ways. To do this, they used two different ways of mapping the data. The first method, known as planar density estimation, creates a smooth, heat-map-like view of the city, showing broad areas of high activity. The second method, network density estimation, is more precise; it calculates intensity only along the actual lines of the street network, respecting the fact that people and buildings are connected by roads, not by straight lines through the air. By comparing these two approaches, the team could see both the big picture and the fine details of how Airbnb and urban renewal were reshaping Lisbon.

The broad view revealed a clear pattern: both Airbnb listings and building renovations were heavily concentrated in the historic center of Lisbon and its immediate surroundings. This central belt, which includes famous areas like Baixa, Chiado, and Alfama, showed a steep drop-off in activity as one moved away from the core. The two phenomena seemed to share the same territory, suggesting that the areas most attractive for tourists were also the areas receiving the most investment in building restoration. However, this flat map view, while useful for seeing the general shape of the city, smoothed over important details. It made the central area look like a single, uniform patch of high activity, hiding the fact that the city is actually made up of distinct streets and blocks.

When the researchers switched to the network-based view, the picture changed significantly. The broad central patch broke apart into specific, linear corridors. Instead of a vague zone of activity, the data revealed that both Airbnb and urban renewal were tightly organized along particular streets and connected at specific junctions. The study identified five main systems where this concentration was strongest, including the famous Avenida da Liberdade and the narrow, winding streets of the historic quarters. In these network maps, it became clear that the two phenomena were not just sharing a neighborhood; they were often occupying the exact same street segments or running parallel to each other on adjacent blocks. This level of detail is crucial because it shows that the pressure on the city is not spread evenly across a district but is channeled through specific arteries.

The study also looked at how these patterns held up when zooming in on smaller areas, such as specific parishes and individual streets. In the parish of Santa Maria Maior, which covers the Baixa and Alfama districts, the network view showed a dense lattice of short streets where both Airbnb and renovations were active. The highest intensity points were found at intersections where multiple streets met, places with high visibility and heavy foot traffic. In other areas, like the Avenida da Liberdade, the data showed a continuous line of activity along the main avenue and its side streets. This confirmed that the relationship between short-term rentals and building renewal was not a random coincidence but a structured phenomenon organized by the city's physical layout. The researchers found that even when they adjusted their data to account for different factors, such as the price of a rental or the type of host, the main pattern remained the same. The central corridors and key nodes stayed consistent, suggesting that the location and connectivity of a street are more important than the specific details of the people or businesses using it.

One of the most important findings of the study is that the way we choose to look at a city changes what we see. The broad, flat maps are excellent for identifying general zones of interest, but they can hide the fact that two things happening in the same neighborhood might actually be separated by a few blocks or a topographical barrier. The network maps, by following the streets, reveal the true anatomy of the city's pressure points. This distinction matters for city planners. If a government wants to manage the impact of tourism and renovation, knowing that the pressure is concentrated on specific streets rather than whole neighborhoods allows for much more precise action. Instead of applying broad rules to an entire district, officials can target specific corridors or intersections where the overlap is most intense.

The study does not claim to prove that one thing caused the other. It does not say that Airbnb listings forced building renovations, or that renovated buildings automatically attracted more tourists. Instead, it offers a diagnostic map, showing where these two powerful forces are meeting and how they are arranged in space. The researchers suggest that this combined approach—using both broad and network-based views—provides a much clearer tool for understanding complex urban changes. By identifying the specific streets and nodes that carry the most weight, cities can move from general observations to targeted strategies. The work in Lisbon demonstrates that to truly understand how a city is changing, one must look not just at the areas where things happen, but at the lines that connect them. The results show that the future of urban management may depend on recognizing these invisible networks and the specific points where they carry the heaviest load.