Determinants of the Price of Airbnb Accommodations Through a Weighted Spatial Regression Model: A Case of the Autonomous City of Buenos Aires

Abstract

In the context of the global growth of the collaborative economy, Airbnb has established itself as one of the most influential players in the transformation of the tourist accommodation market, especially in the reconfiguration of urban tourist accommodation. This article examines empirically and critically how this platform operates in Buenos Aires, the most visited city in Argentina and one of the main tourist hubs in South America. Based on a database of 17,249 active listings, the price formation of accommodations is analyzed using a comparative methodological approach between a general linear model (GLM) and a geographically weighted regression (GWR) model. While the GLM allows for capturing general patterns, the GWR reveals significant territorial differences, offering a detailed reading of the spatial behavior of prices in the city. The results show that variables such as the capacity of the accommodation, its type (full house), the host’s condition, the users’ ratings and the proximity to strategic points such as the subway or Plaza de Mayo have a significant influence on prices. In addition, it is shown that the influence of these variables varies by neighborhood, confirming that the pricing logic in Airbnb is deeply territorialized. This study not only provides novel empirical evidence for a Latin American city that has been little explored in the international literature, but also offers useful tools for hosts, urban planners and public decision makers. Its main contribution lies in showing that prices respond not only to accommodation attributes, but also to broader spatial inequalities, opening the debate on the effects of Airbnb on housing access and urban management in cities with strained real estate markets. By shedding light on these territorial asymmetries, the study offers valuable insights for public policy and urban governance and contributes directly to the achievement of Sustainable Cities and Communities (SDG 11), while also supporting Industry, Innovation and Infrastructure (SDG 9) and Reduced Inequalities (SDG 10), by providing practical knowledge that fosters more equitable and sustainable urban development.

1. Introduction

In recent years, the collaborative economy has received renewed academic attention, especially because digital platforms have transformed the way tourists find accommodations. Airbnb, in particular, has redefined urban hospitality and raised challenges regarding housing access, regulation, and sustainability [1,2,3,4]. Recent studies show that discussions about these platforms have shifted towards an applied sustainability approach that is less focused on normative declarations and more focused on practical governance, regulatory, and certification mechanisms [5,6,7].

In this sense, the concept of pragmatic sustainability has gained ground. It is understood as the pursuit of sustainable solutions that are adaptable, contextualized, and verifiable in specific urban settings. Recent research demonstrates that environmental policies and private certifications directly impact tourism flows and destination value, providing examples of pragmatic sustainability in territorial management [8,9]. Similarly, Airbnb’s presence has been shown to influence urban conservation and rental price dynamics, underscoring the importance of integrating pragmatic approaches into city regulations [10]. However, this model has also generated intense debates around its impacts on housing, tourism, informal employment and urban regulation [11].

Furthermore, there is a growing sensitivity to sustainability in tourism demand, particularly among younger tourists. These tourists value certifications, digital filters, and verifiable practices that ensure coherence between tourism consumption and environmental responsibility [12]. This idea is reinforced by the notion that, when approached pragmatically, sustainability not only responds to regulatory requirements, but also constitutes a factor of competitiveness and legitimacy for destinations and digital platforms.

In academia, a substantial part of the literature has focused on analyzing the determinants of the price of Airbnb listings, understanding that price not only conditions guest choice, but also directly affects host profitability and, therefore, market dynamics [13,14]. Traditionally, these studies have employed global regression models that, although useful, do not consider the possible spatial variation in the effects of the explanatory variables, which represents an important limitation in urban contexts with high territorial heterogeneity.

Within this framework, the present study analyzes the determinants of Airbnb accommodation prices in the Autonomous City of Buenos Aires (CABA) by comparing a general linear model (GLM) and a geographically weighted regression (GWR) model. Through this methodological approach, the study seeks to answer two central questions: Which spatial and reputational factors influence Airbnb pricing in Buenos Aires, and how do these relationships vary across the urban landscape?

The remainder of this paper is organized as follows: Section 2 discusses previous research, Section 3 presents the study area and data, Section 4 explains the methods used in this paper and Section 5 details the results. Finally, Section 6 presents the conclusions and future lines of research.

2. Literature Review

Since 2008, Airbnb’s rapid growth has sparked extensive academic debate about its benefits, challenges, and regulatory frameworks. While the platform is recognized for offering competitive advantages such as lower costs and more authentic experiences for users, it also requires clear public policies to ensure its legal and sustainable operation [7,15,16].

The platform’s impact on the hotel and tourism industry has been the subject of numerous studies. Some evidence shows that Airbnb’s entry reduced hotel revenues in certain contexts, while in others, it generated employment and diversified the tourism offerings. However, the marginal effects diminish as the platform grows [17,18,19,20].

Regarding the housing cost, the proliferation of short-stay listings has been found to increase purchase prices and rents by shifting units from the residential to the tourist market [21,22,23]. However, these effects show clear spatial variability, calling into question the usefulness of global models and highlighting the need for approaches sensitive to territorial heterogeneity [24,25,26].

Pricing also depends on trust- and reputation-related factors [27]. Signals such as the average rating, the volume and length of reviews, the age of listings, and Superhost status reduce information asymmetry and influence the decisions of guests and hosts [28,29,30]. Physical and location attributes, such as the number of rooms, proximity to points of interest, and level of services, also influence pricing, although the effect varies depending on the urban context [31,32,33].

From another perspective, numerous authors have analyzed the role of trust and reputation in shaping user behavior on platforms like Airbnb. Several studies highlight how guest experience and service quality influence the platform’s commercial reputation [34,35,36]. Key factors such as “Superhost” status [37,38], perceived authenticity, pricing strategies, and brand personality play a significant role in shaping consumer perceptions [39].

In this context, previous research has examined Airbnb’s evaluation system and found that the Superhost designation—granted to hosts with a minimum rating of 4.8, consistent booking activity, and no cancellations—is positively associated with higher guest satisfaction and booking frequency [40]. This status, in turn, enhances the host’s revenue potential.

In the case of Argentina, and particularly in the Autonomous City of Buenos Aires (CABA), evidence remains limited. Available studies have focused on traditional variables—such as number of rooms, type of accommodation, urban location, user ratings, and certain socioeconomic indicators—but they tend to rely on global models that assume spatial homogeneity, which limits their ability to capture the territorial diversity characteristic of such a heterogeneous city [41,42].

With the aim of overcoming these limitations, the most recent literature has incorporated geographically weighted regression (GWR) approaches and machine learning techniques. These studies show differentiated spatial patterns in cities such as Los Angeles and New York [23], show significant variations in the impact of density in Sydney [43], and confirm the superiority of GWR over global models in contexts such as the United Kingdom [24].

The application of more sophisticated models has spread across Europe and North America in ten major European cities, with size, location, and quality of accommodation established as the main factors determining price [19]. At the same time, methodologies based on neural networks and explainable machine learning have made it possible to incorporate advertising texts, points of interest, and statistical data to improve price prediction in cities such as Austin and Dublin [6,44]. Finally, normative tests suggest that measures such as 180-day limits on short-term rentals can reduce long-term rental prices by 2% to 3%, demonstrating how public policies directly affect the performance of these platforms [7].

Despite the growing body of literature on Airbnb pricing, spatial modeling techniques such as Geographically Weighted Regression (GWR) remain underutilized—particularly in Latin American contexts. This study addresses that gap by offering a dual contribution: methodologically, through the application of GWR, and contextually, by focusing on Buenos Aires—a city largely underrepresented in the global literature. In doing so, it enhances our understanding of how spatial price determinants operate in platform economies embedded within heterogeneous urban environments.

Beyond the technical dimension, the recent literature introduces the notion of pragmatic sustainability, understood as an adaptive, results-oriented approach that connects the regulation of digital platforms with sustainable development goals. Within this framework, Airbnb can either promote or hinder the Sustainable Development Goals—particularly SDGs 9, 10, and 11—depending on how it is integrated into urban and housing policies [5,45,46].

Furthermore, this research contributes to the emerging literature that connects platform economies with the Sustainable Development Goals (SDGs), a dimension that remains largely overlooked in Airbnb studies. While previous research has advanced our understanding of pricing mechanisms and spatial impacts, few studies have explicitly examined their alignment with global sustainability agendas. Recent contributions emphasize the relevance of short-term rental platforms in shaping urban futures aligned with SDG targets. For example, Alburquerque et al. [5] argue that platforms like Airbnb can either support or hinder progress toward Sustainable Cities and Communities (SDG 11), depending on how they are regulated and integrated into broader urban policy frameworks. Likewise, Bei and Celata [45] highlight the socio-spatial consequences of digital tourism, noting its capacity to either reproduce or reduce urban inequalities, thus linking to Reduced Inequalities (SDG 10).

In parallel, Airbnb has also been analyzed as a driver of digital innovation and infrastructure, aligning with the goals of Industry, Innovation and Infrastructure (SDG 9). In this regard, Barrera-Martínez and Parra-López [46] examine the platform’s transformative role in tourism through digitalization and organizational innovation, arguing that its business model constitutes a new form of digital infrastructure that enables novel economic interactions.

By incorporating these perspectives, this study not only contributes to the methodological and contextual literature on Airbnb but also expands the discussion by situating its findings within the broader framework of sustainable urban development.

3. Materials and Methods

3.1. Study Area and Data

Argentina, with a population of 47.07 million inhabitants according to INDEC [47], is located at the southern tip of the American continent. In recent years, tourism in the country has experienced remarkable growth, with a special highlight in 2019, when an all-time record was reached in the number of annual visits. According to INDEC figures (2025), the interannual variation was +11%, registering 2.7 million non-resident tourists who entered by air. If arrivals by sea and land are also considered, this figure doubles.

The city of Buenos Aires is by far the country’s main tourist destination. During the first half of 2019, 1,465,025 international tourists were registered, and that figure was projected to reach three million by the end of the year. This tourism boom was accompanied by an expansion of the flow of visitors to less traditional neighborhoods, which generated a broader economic and social impact on the porteño territory.

In 2019, Buenos Aires consolidated its position as the most visited city in South America and was included among the 100 most visited cities in the world, according to a report by Euromonitor International. Considered the “most European destination in the Americas,” the Argentine capital combines an outstanding gastronomic offer with a vibrant artistic, cultural and sports scene. At the national level, no other city concentrates such a significant volume of tourists as Buenos Aires. In recent years, it has also positioned itself as a key center for gastronomy, sports, culture and business tourism. As the Federal Capital, it has first class infrastructure and tourist services, with a wide range of hotels, restaurants, theme parks, museums, cultural centers, stadiums and theaters distributed in its different neighborhoods.

As of September 2022, the Airbnb platform registered 17,249 active offers in the City of Buenos Aires. Of these, 15,396 corresponded to complete homes, 1979 to private rooms, 119 to hotel rooms and 192 to shared rooms. As for their location, almost half of the 17,249 units were concentrated in two neighborhoods: 6086 in Palermo and 2754 in Recoleta. In addition, 14,004 of these offers were located in Communes 1, 2, 13 and 14, which means that three quarters of the total offer is located in the northern corridor, the area with the highest real estate appreciation in the city.

The analysis presented in this study is based on a set of data extracted from ads published on the Airbnb platform [48] (www.airbnb.com.ar) corresponding to the Autonomous City of Buenos Aires during the months of October and November 2022 (Figure 1).

3.2. Methodology

To identify the factors that influence the prices of accommodations offered through the Airbnb platform in the Autonomous City of Buenos Aires, two complementary econometric approaches were used: the General Linear Model (GLM) and the Geographically Weighted Regression (GWR). While the GLM allows capturing global relationships between the explanatory variables and the dependent variable (price of lodging), under the assumption of spatial homogeneity in the estimated parameters, the GWR model introduces greater flexibility by allowing these parameters to vary locally according to geographic position. This adaptability of the GWR is particularly useful in urban environments with high territorial heterogeneity, as is the case of Buenos Aires, where the characteristics of the urban environment can substantially modify the impact of certain attributes on price. The technical specifications of both models and the theoretical foundations that justify their use in this study are presented below.

3.3. General Linear Model (GLM)

The General Linear Model is one of the most widely used models because of its interpretative simplicity and its ability to estimate global effects, i.e., effects that are assumed to be constant throughout all observations, without considering spatial or temporal variations.

In the GLM, the dependent variable is estimated with a set of dependent variables globally. The model specification is expressed as:

$$y_i = {\beta }_0 + \sum {\beta }_k X_{ik} + {\epsilon }_i$$

(1)

where

• y_i represents the i-th observation of the dependent variable (in this case, the quoted price of the accommodation on Airbnb).
• β₀ is the intercept or constant term of the model
• β_k are the estimated coefficients associated with each explanatory variable X_ik
• X_ik represents the value of the k-th independent variable for observation i
• ε_i is the random error term that captures the effects not explained by the model.

The GLM assumes that the relationships between the explanatory variables and the dependent variable are linear and homogeneous throughout the analyzed space. In addition, the model requires the fulfillment of certain classic assumptions of linear regression, such as normality of the error, homoscedasticity (constant error variance), absence of autocorrelation and non-multicollinearity between the explanatory variables.

3.4. Geographically Weighted Regression

Geographically Weighted Regression (GWR) is a spatial modeling technique that allows us to explore how the relationships between a dependent variable and a set of explanatory variables vary in geographic space. Unlike global linear models, which assume constant relationships between variables throughout the study domain, GWR estimates local coefficients for each observation, which allows capturing the spatial heterogeneity of the analyzed phenomenon.

The general formulation of the GWR model is as follows Fotheringham et al. [49].

$$y_i = {\beta }_0(u_i, v_i)+{\sum }_{k=1}^m{\mathsf{\beta }}_{\mathrm{k}\mathrm{i}}(u_i, v_i)X_{ik} + {\epsilon }_i$$

(2)

where

• y_i is the dependent variable at location i,
• X_ik is the value of the k-th explanatory variable in the i-th observation,
• β_k (u_i, v_i) are the estimated local coefficients for the variable k at location (u_i, v_i),
• ε_i is the random error term associated with observation i.

The model estimates a different set of parameters $\widehat{\beta }$ for each geographic point, using a weighted least squares approach. The local coefficients are calculated using the following matrix expression:

$$\widehat{\beta }\left(i\right)={\left[X^T{W}_{\left(i\right)}X\right]}^{-1}{X}^T{W}_{\left(i\right)}Y$$

(3)

where W_(i) indicates a spatially weighted diagonal matrix. X indicates the data matrix of the explanatory variables. Y is the vector of the dependent variable. In matrix form these variables are formulated in matrices (4)–(6).

$$W_{\left(i\right)}=\left[\begin{array}{cccc}{W}^{i1}& 0& \dots & 0\\ 0& W^{i2}& \dots & 0\\ ⋮& ⋮& ⋮& ⋮\\ 0& W^{in}& \dots & 0\end{array}\right]$$

(4)

$$X=\left[\begin{array}{cccc}1& x_{11}& \dots & x_{m1}\\ 1& x_{12}& \dots & x_{m2}\\ ⋮& ⋮& ⋮& ⋮\\ 1& x_{1n}& \dots & x_{mn}\end{array}\right]$$

(5)

$$Y=\left[\begin{array}{c}{y}_1\\ y_2\\ y_3\\ y_4\end{array}\right]$$

(6)

The assignment of weights in W_(i) is based on spatial decay functions, which give greater relevance to observations close to point i and less to those farther away. In this study, a Gaussian weighting function is used, defined as:

$$W_i\left(y_0\right)=\exp (-0.5{\left({\displaystyle \frac{d{y}_{0}j}{b}}\right)}^2)$$

(7)

where

• dy₀j indicates the distance between location i and observed point j,
• b indicates the distance bandwidth which. A given distance or a fixed number of neighborhoods can be used to define the bandwidth.

The optimal bandwidth value is determined by a calibration process that minimizes the Akaike Information Criterion (AIC), selecting the model with the highest explanatory power and lowest relative complexity.

4. Results and Discussion

The analysis with the GLM made it possible to identify the main factors that globally influence the price determination of the accommodations published on the Airbnb platform in the Autonomous City of Buenos Aires. The variables considered included physical attributes of the accommodation (capacity, number of rooms, type of unit), host characteristics (professionalization, seniority), reputational metrics (number and length of reviews, average rating), and spatial variables (distances to landmarks and belonging to specific neighborhoods such as Palermo or Recoleta).

The results, presented in

Table 1. Results of the General Linear Model (GLM).

Variable	Coeffient	p-Value
Constant	22.4299	0.1696
accommodat	0.3976 **	0.0498
bedrooms	0.8734	0.4185
rating	1.2772 ***	0.0006
number_reviews	−0.0144	0.0253
large_reviews	−0.0154 **	0.0143
professional_host	−1.7714 **	0.0500
age	0.7171 ***	0.0073
whole_house	0.5932 **	0.0392
Palermo	0.3092 **	0.0268
Recoleta	0.0634 **	0.0247
Dist_Pink_House	0.0004	0.7801
Dist_May_Plaza_May	−0.0004 **	0.0285
dist_min_meter	0.0003 **	0.0183

Source: own elaboration with data from inside Airbnb. Signif. codes: ‘***’ 0.001 ‘**’ 0.05.

, indicate that variables such as the capacity of the accommodation (accomodat), whether it is an entire house, how long the house has been listed on Airbnb (age) and the average rating (rating) have positive and significant effects on the price. In contrast, the number (number_reviews) and length of reviews (large reviews), as well as the distance to the nearest subway (dist_min_metro), and to Plaza de Mayo (Dist_Plaza_Mayo), show a statistically significant negative relationship.

For example, a one-unit increase in accommodation capacity is associated with a 39% increase in price, while being a full unit increases the price by 71%. The average rating also shows a robust effect: each additional point in rating translates into a 121% increase in quoted price. In the opposite direction, an increase in the number of reviews is associated with a 1.4% decrease in price, and a higher number of characters in the reviews reduces the price by approximately 1.5%, which could suggest a deliberate strategy of lower prices to capture higher review volume.

Regarding spatial variables, it was observed that a greater distance to the nearest subway station or to Plaza de Mayo reduces the price, which evidences the valuation of accessibility and proximity to areas of historical and tourist interest.

The GWR model refines the analysis by capturing spatial variability in the estimated coefficients. Unlike the global model, the GWR estimates a set of localized coefficients for each observation, allowing for the capturing of intra-urban heterogeneity in the impact of each factor

Table 2. Geographically Weighted Model Results (GWR).

Variable	Min.	1st Qu.	Median	3rd Qu.	Max.
Constant	21.4609	22.2913	22.4817	22.8184	23.5378
accommodat	0.3576	0.3734	0.3854	0.4084	0.4324
Bedrooms	0.8040	0.8502	0.8950	0.9189	0.9499
Rating	1.2384	1.2523	1.2596	1.2764	1.2935
number_reviews	−0.0148	−0.0146	−0.0144	−0.0141	−0.0140
large_reviews	−0.0162	−0.0156	−0.0150	−0.0147	−0.0143
professional_host	−1.8223	−1.7946	−1.7708	−1.7440	−1.6959
Age	0.6754	0.6906	0.7054	0.7294	0.7508
whole_house	0.4080	0.5502	0.6880	0.7672	0.8627
Palermo	0.2984	0.3017	0.3045	0.3087	0.3127
Recoleta	0.0543	0.0579	0.0595	0.0637	0.0681
Dist_Pink_House	0.0003	0.0004	0.0004	0.0005	0.0005
Dist_May_Plaza_May	−0.0006	−0.0005	−0.0005	−0.0004	−0.0004
dist_min_meter	−0.0003	−0.0004	−0.0004	−0.0004	−0.0004

Source: own elaboration with data from inside airbnb.

summarizes the statistical values of the estimated coefficients for each variable, including the minimum value, lower quartile, median, upper quartile and maximum value. This statistical variation is evidence that the relationships between factors and price are not constant across the territory but depend on the specific spatial context.

For example, the lodging capacity variable shows an average coefficient close to 0.39, with a range from 0.36 to 0.43. This means that, on average, a one-unit increase in lodging capacity increases the price by about 39%, although this effect may be more or less pronounced depending on the area. Similarly, the whole house attribute had a moderately positive effect (median = 0.688), but its coefficient varied between 0.41 and 0.86, suggesting a strong influence of the territorial context on the valuation of this attribute.

In the case of the average rating variable, the coefficient showed high spatial consistency (narrow range between 1.24 and 1.29), but its interpretation becomes more complex when negative correlations are observed in certain neighborhoods. This could be related to guest expectations in areas with low prices, which generate more lenient ratings.

Reputation-related variables also showed interesting results. The coefficient of number of reviews is negative and consistent across the territory (average ≈ −0.0144), indicating that listings with a higher volume of reviews tend to have lower prices, possibly as part of a competitive price positioning strategy. The length of the reviews also presented a negative relationship (average ≈ −0.0150), which reinforces the hypothesis that the most commented listings are the cheapest, perhaps due to their high turnover and volume of bookings.

Regarding spatial variables, it was observed that the distance to Plaza de Mayo has a negative coefficient in most cases (median = −0.0005), indicating that accommodations closer to this historical point present higher prices. On the other hand, the distance to the nearest subway also shows a mostly negative relationship (median = −0.0004), suggesting that accessibility to public transportation is positively valued by guests. In both cases, the magnitude of the effect decreases in peripheral neighborhoods, reflecting a spatial elasticity in the valuation of these attributes.

Finally, the negative coefficient of the professional host (average = −1.7708) stands out, whose magnitude varies considerably between neighborhoods (−1.82 to −1.69), suggesting a moderate penalty or a strategy of lower prices to compete with particular hosts, possibly in response to market saturation in some areas.

To complement the quantitative results presented in Table 2, thematic maps were produced (Figure 2, Figure 3 and Figure 4) that represent the spatial distribution of the coefficients estimated by the GWR model for some of the most relevant variables: distance to the nearest subway, age of the listing, number of reviews, and average rating. These maps allow us to identify specific spatial patterns, revealing areas where certain variables exert greater or lesser influence on the listing price of Airbnb listings.

Figure 2 showed that, in the central areas of the city, particularly in the neighborhoods of Recoleta, San Nicolás and Monserrat, proximity to subway stations has a more pronounced and negative impact on price. That is, listings located within walking distance of public transportation had significantly higher prices, confirming tourists’ valuation of accessibility. As accommodations are further away from the urban core, this effect tends to be diluted, reflecting a spatial elasticity in the valuation of this attribute. The t-values indicate that this relationship is statistically significant in most neighborhoods in the center and north of the city.

In Figure 3, the age of the listing on the platform also showed a positive relationship with price, especially in central and high tourist density areas. In neighborhoods such as Palermo, Retiro and Recoleta, older listings tend to set higher prices. This situation could be attributed to the experience accumulated by the hosts, their consolidated reputation and a greater optimization of commercial strategies. This variable acts as a proxy for host maturity on the platform, and its positive effect on price is spatially heterogeneous, but statistically robust in the northern corridor of the city.

Contrary to what is reported in most of the international literature, the results shown in Figure 4a indicate a negative effect of the number of reviews on price. In most of the neighborhoods analyzed, especially in intermediate and peripheral areas, a higher number of reviews is associated with lower prices. This suggests a market strategy in which hosts reduce their rates to attract higher volume of guests and thus increase their presence and visibility within the platform. The t-values confirm the statistical significance of this relationship in most of the territory.

Figure 4b shows that the relationship between guest ratings and price is ambiguous. In some neighborhoods in the center and north, the coefficients are slightly positive, while in other areas they are negative or close to zero. This inconsistency could be explained by the high concentration of positive ratings that characterize Airbnb, a phenomenon widely documented in previous studies. The high homogeneity in ratings reduces their discriminant capacity and weakens their effect on price formation. Consequently, although statistically significant in some areas, the average rating does not seem to be a robust predictor of price in Buenos Aires.

These spatial variations can be better understood when considering the socio-economic contrasts of Buenos Aires. The city exhibits pronounced real estate asymmetries between its northern corridor—wealthier, with higher land value—and southern areas, characterized by lower infrastructure development and reduced tourist demand. This duality explains why attributes such as accessibility, host reputation, or even professionalization yield different price effects across neighborhoods.

Comparison Between the GWR and GLMs

The results indicate that the GWR model far outperforms the GLM in terms of explanatory power and statistical fit. The coefficient of determination (R²) of the GWR model was 0.69 versus 0.58 for the GLM, while the adjusted R² was 0.64 versus 0.56, respectively. In addition, the GWR model presented a lower value of Akaike’s Information Criterion (AIC = 514.997) versus the GLM (AIC = 532.64), confirming its superiority in terms of fit and parsimony (

Table 3. Performance comparison between GLM and GWR models.

Performance	Model	Model GWR
R2	0.58	0.69
Adjusted R2	0.56	0.64
AIC	532.64	514.997
Number of parameters	13	13

Source: own elaboration with data from inside airbnb.

).

The results highlight the relevance of considering the spatial dimension when modeling prices in heterogeneous urban markets. While the GLM provides a useful global view as a baseline, the GWR allows capturing territorial differences crucial to understand the logic of price formation in platforms such as Airbnb.

5. Conclusions and Policy Implications

This study analyzed the determinants of Airbnb accommodation prices in the Autonomous City of Buenos Aires, integrating physical and reputational attributes with a spatial analytical perspective. Through a comparative methodological approach—using both a General Linear Model (GLM) and a Geographically Weighted Regression (GWR)—it demonstrates that price formation is influenced not only by listing characteristics but also by their spatial location within the urban structure.

The results confirm that variables such as accommodation capacity, unit type (entire property), host professionalization, listing age, user ratings, and proximity to urban landmarks significantly affect pricing. While these findings are consistent with prior studies (e.g., [37]), they also highlight unique aspects of the Buenos Aires context. Notably, the negative relationship between the number of reviews and price contradicts international trends and may reflect local strategies aimed at improving platform visibility through lower initial rates.

The use of the GWR model proved particularly valuable in capturing strong spatial heterogeneity. In central neighborhoods like Palermo or Recoleta, certain attributes positively influence price, whereas in peripheral areas their effects are diminished or even reversed. These findings underscore that Airbnb operates within a fragmented and spatially uneven market, shaped by the socio-spatial dynamics of the city.

The integration of pragmatic sustainability provides an additional conceptual lens for interpreting these results. Rather than approaching sustainability as an abstract or normative goal, pragmatic sustainability emphasizes context-specific, verifiable, and actionable solutions. In the case of Buenos Aires, this means adopting flexible regulations that safeguard housing affordability, promote equitable access to urban opportunities, and preserve cultural heritage, while still allowing the collaborative economy to generate income and stimulate innovation. By linking practical regulatory tools with broader sustainability objectives, pragmatic sustainability bridges the gap between theory and practice, ensuring that the benefits of platforms like Airbnb are distributed more fairly while reducing their negative externalities on urban housing and communities.

From a theoretical standpoint, this study provides empirical evidence of spatial non-stationarity in urban short-term rental markets, contributing to the literature on the collaborative economy, digital platforms, and urban geography. Methodologically, the application of GWR represents a significant innovation, offering a more nuanced and context-sensitive analytical tool that overcomes the limitations of conventional global models.

From a policy perspective, these findings emphasize the importance of differentiated regulations that reflect the heterogeneous impacts of short-term rentals across neighborhoods. Areas with strong tourism demand, high property values, and limited housing supply are particularly vulnerable to distortions in long-term rental markets. Regulatory tools such as rental caps, progressive taxation, and mandatory registration schemes can help mitigate displacement effects, while targeted incentives may encourage hosts to adopt sustainable practices, including energy efficiency measures, fair pricing strategies, and compliance with housing standards. Local governments could also use spatial models, such as GWR, to identify critical areas where interventions are most urgent, ensuring that regulatory responses are not only effective but also geographically equitable.

6. Future Research

Despite its contributions, this study has limitations. The analysis was based on cross-sectional data from a single period, restricting the ability to capture seasonal fluctuations or temporal changes in Airbnb dynamics. In addition, the absence of neighborhood-level socioeconomic indicators may limit the explanatory power of the models. Future research could expand this work by incorporating temporal dimensions, exploring other Latin American cities for comparative purposes, and integrating additional variables such as property amenities, host management strategies, or guest demographic profiles. Longitudinal analyses could also provide insights into how short-term rental markets respond to external shocks, including regulatory reforms, economic crises, or global events such as the COVID-19 pandemic. Such extensions would enrich the understanding of how collaborative platforms evolve and interact with broader urban systems.

Advancing this interdisciplinary research agenda at the intersection of urban economics, tourism studies, geography and public policy will enable future studies to provide a more comprehensive, contextually grounded understanding of the territorial impacts of digital accommodation platforms in contemporary cities.

Finally, these findings should be interpreted in light of Buenos Aires’ unique socio-spatial configuration, which combines areas of high tourist infrastructure with others characterized by lower development and demand. This duality helps explain why factors such as accessibility, host reputation, or professionalization exhibit varying effects on pricing across different neighborhoods.

In alignment with the 2030 Agenda for Sustainable Development, this study offers evidence-based insights that can inform policies supporting more sustainable, inclusive, and resilient urban communities (SDG 11). By revealing the spatial disparities in short-term rental pricing and platform dynamics, the study contributes to identifying areas of intervention that can help reduce urban inequalities (SDG 10) and optimize digital infrastructure for equitable growth (SDG 9). Furthermore, the findings advance the concept of sustainable communities, understood not only as environmentally resilient, but also as socially just and spatially balanced urban environments. By adopting a pragmatic approach—grounded in localized data and spatial analysis— the research proposes tools for public and private stakeholders to make informed decisions that benefit both hosts and residents, contributing to more equitable access to housing and urban opportunities in the digital tourism era.