Spatial Modelling of Urban Accessibility: Insights from Belgrade, Republic of Serbia

Abstract

This study presents the first comprehensive spatial accessibility assessment of essential urban services in Belgrade, Republic of Serbia, conducted entirely with open-source tools and data. The analysis focused on six facility categories: primary healthcare centers, public pharmacies, primary and secondary schools, libraries, and green markets. Spatial accessibility was modelled using OpenRouteService (ORS) isochrones for walking travel times of 5, 10, and 15 min, combined with population data from the Global Human Settlement Layer (GHSL). Results indicate that 79% of residents live within a 15-min walk of a healthcare facility, 74% of a pharmacy, 89% of an elementary school, 52% of a high school, 60% of a library, and 62% of a green market. Central administrative units such as Vračar, Zvezdara, and Stari Grad demonstrated nearly complete service coverage, while peripheral areas, including Resnik, Jajinci, and Višnjica, exhibited substantial accessibility deficits, often coinciding with lower-income zones. The developed workflow provides a transparent, replicable approach for identifying underserved neighborhoods and prioritizing investments in public infrastructure. This research emphasizes the role of spatial accessibility analysis in advancing Sustainable Development Goals (SDGs), contributing to the creation of more inclusive, walkable, and sustainable urban environments, while on the other hand, it offers practical insights for improving urban equity, guiding policy formulation, and supporting necessary planning decisions. Subsequent research will focus on alternative facilities, other cities such as Novi Sad and Niš, and the disparity between urban and rural populations.

1. Introduction

In the modern age of urbanization, sustainable urban development has emerged as an essential concept for establishing livable and equitable cities. The expansion of cities presents a significant challenge in ensuring that essential services of the modern lifestyle are accessible to all residents, which directly affects quality of life, social equity, and environmental sustainability. Belgrade, the capital of the Republic of Serbia, reflects this global trend.

Belgrade’s urban trajectory has been shaped by repeated shifts between ambitious urban plans and partial implementation, with recent decades characterized by strong peripheral expansion. This growth has not always been matched by adequate infrastructure or public service provision, and constraints such as governance fragmentation, delayed investment in facilities, and disparities in service distribution have reinforced differences between central and outer neighborhoods. These dynamics are directly reflected in the accessibility of everyday services, making the evaluation of walking access to healthcare, schools, libraries, and markets particularly relevant for identifying inequalities and informing strategies for more balanced urban development.

Therefore, this paper analyzes the spatial accessibility of essential urban services, including primary healthcare facilities, pharmacies, primary and secondary educational institutions, libraries, and green markets, within the densely urbanized region of Belgrade, Republic of Serbia, with the objective of offering insights that foster the development of an inclusive, sustainable, and equitable city.

Access to essential services is crucial for the achievement of various Sustainable Development Goals (SDGs), such as eradicating poverty (SDG 1), ensuring affordable healthcare (SDG 3), providing quality education (SDG 4), promoting gender equality (SDG 5), reducing inequalities (SDG 10), and fostering sustainable urban development (SDG 11) [1]. A city that provides residents with convenient access to educational, health, and cultural facilities, alongside essential goods and services, increases social inclusion, mitigates disparities, and cultivates a pedestrian-friendly atmosphere that supports environmental sustainability [2,3]. Given the significance of these objectives, there is an urgent need for comprehensive studies in the Republic of Serbia that simultaneously investigate the spatial accessibility of various urban services, resulting in a notable deficiency in understanding and policy formulation.

While accessibility analyses are increasingly present in Central and Eastern Europe, most focus on single sectors such as healthcare or education. For instance, Romanian studies have examined hospital access [4,5,6] and high school accessibility [7], while others addressed cardiovascular care [8] or broader disparities in service provision [9]. Another similar research that tackles the topic of walkability is found in Albania, where the research has focused on school travel behavior in Tirana [10]. However, to date, there is no study in the Balkans that systematically integrates multiple categories of essential services, such as healthcare, pharmacies, schools, libraries, and green markets, into a single framework of accessibility assessment. This paper, therefore, provides a novel and strategic analysis of multi-service accessibility in Belgrade.

This paper enhances the current research landscape by providing a thorough analysis encompassing various objects of interest, in contrast to prior studies that generally concentrate on a singular service type. This research offers a comprehensive view on the accessibility of essential urban services by incorporating the spatial analysis of primary and secondary educational institutions, primary healthcare facilities, pharmacies, libraries, and green markets. This study is the first research of its kind in the Republic of Serbia, filling a significant research gap in spatial planning research in the Balkan region. In the present study, a methodology based entirely on open-source and free software was utilized, ensuring transparency, replicability, and adaptability for other cities or regions. The findings seek to offer practical recommendations for equitable distribution of urban services in Belgrade. Areas with limited accessibility to the aforementioned services were identified, and specific measures that promote a more inclusive and pedestrian-friendly environment were proposed. These measures correspond with the principles of sustainable development, highlighting affordable healthcare, accessible education, diminished inequalities, and the establishment of inclusive and resilient urban environments. These results are aimed to be a valuable resource for urban planners, policymakers, researchers, and all parties related to spatial analysis problems, in general, who are dedicated to creating cities that prioritize the well-being of all residents.

Spatial accessibility is vital in determining the availability and fair distribution of essential services, including healthcare, education, and recreational facilities. Notably, a significant aspect of spatial accessibility is that pedestrian access to the aforementioned attractions positively influences property values and the economic development of a specific area [11]. A variety of studies have investigated the impact of spatial and temporal variations on access to these services, especially in urban and rural contexts, emphasizing the significance of demand, supply, and transportation conditions [12,13,14,15,16,17,18]. Research in healthcare has highlighted the significance of incorporating spatiotemporal factors, including variations in demand, supply, and traffic conditions, to achieve a dynamic understanding of service delivery. For example, research in Seoul, South Korea, has employed this methodology to evaluate the temporal and spatial variations in healthcare access, mirroring actual shifts in service availability [19]. This dynamic approach offers a more accurate representation of accessibility, considering peak demand intervals and traffic variations that may otherwise be neglected. Conversely, these models may be overly complex for policymakers to implement and may inadequately address all dimensions of social and economic inequality, which can also influence accessibility. Research in China has revealed substantial inequalities in spatial access to healthcare, particularly in rural and remote regions, which face significant accessibility challenges [20,21,22,23]. These findings underscore the persistent urban–rural disparity in healthcare access; however, they occasionally neglect to consider local contextual elements, such as regional policies or the influence of informal healthcare providers in reducing accessibility challenges. The distribution and clustering of healthcare centers are critical factors in understanding these disparities, as evidenced by research conducted in Beijing and Tehran [24,25].

Accessibility studies have focused on educational services, especially in rural regions, with the objective of improving equity in education. Research in Changyuan, China, has highlighted the necessity of optimizing school placement and service coverage to guarantee equitable access for students, particularly in underserved regions [26]. Furthermore, research on early childhood education in Melbourne, Australia, underscores the impact of geographic location on educational quality while also questioning the metrics used for quality assessment and the sufficiency of measures addressing geographic disparities in quality [27]. These studies underscore the necessity for spatial optimization of educational facilities to enhance service coverage and guarantee that all students have access to quality education, irrespective of their geographic location [28].

Beyond healthcare and education, research has examined the spatial accessibility of various public services, including pharmacies, libraries, and recreational facilities [29,30,31]. Research conducted in Detroit, USA, and Kenya examined the impact of travel distances and transportation availability on access to healthcare services and pharmacies, highlighting the essential role of mobility in spatial accessibility [32,33]. Research conducted on libraries in Florida, USA, and Hong Kong has employed Geographical Information Systems (GIS) and spatial analysis to evaluate the accessibility of public libraries, highlighting the influence of geographic and demographic variables on equitable access to these services [34,35].

Transportation is essential for affecting spatial accessibility, especially in urban settings where travel distances impact access to diverse services. Research indicates that restricted or absent transportation intensifies disparities in access to healthcare, education, and recreational services. This phenomenon has been documented in research papers investigating the influence of travel distances on accessibility to educational institutions and other recreational facilities [27,32]. The promotion of modern urban planning is thus advocated, encompassing the design of a pedestrian-friendly city, often referred to as a 5-, 10-, or 15-minute city [36]. The idea of the 15-minute city provides an important conceptual frame for this research. This concept envisions urban environments where residents can meet their daily needs: work, education, healthcare, shopping, and recreation, within a short walk or bike ride [37]. By reducing car dependency, this model emphasizes equity, sustainability, and human-scale development. Assessing Belgrade’s service distribution through 15-min walking isochrones thus directly links local urban realities with broader international debates on liveable and sustainable cities.

GIS and spatial analysis tools are crucial in determining spatial accessibility [38,39]. These tools are employed to model and assess service distribution, facilitating the identification of regions with inequitable access. The floating catchment area method has been employed to evaluate the spatial accessibility of libraries in the USA and Hong Kong, offering an alternative methodological framework for analyzing the serviceability of specific regions [34,35]. GIS-based models may explain spatial disparities in service provision, thereby assisting urban planners and policymakers in addressing these challenges more effectively.

This paper presents an application of spatial accessibility analysis in a high-density area of Belgrade, Republic of Serbia, with the following objectives:

• To conduct a comprehensive spatial analysis of multiple facilities, including primary and secondary educational institutions, primary healthcare centers, libraries, pharmacies, and green markets.
• To provide direct policy recommendations for decision-makers in the research area of Belgrade regarding underrepresented areas and the necessity for further investments to enhance the welfare of residents in these regions.
• To provide a diverse group of researchers, governmental professionals, and others a direct workflow aimed at enabling these groups to conduct spatial accessibility analyses utilizing free and open-source tools.

The remainder of this paper is structured as follows. Section 2 outlines the study area, data sources, and methodological framework. Section 3 presents the results of accessibility modelling across different services. Section 4 discusses policy implications, limitations, and directions for future research. Finally, Section 5 concludes by emphasizing the broader relevance of accessibility analysis for equitable urban development.

2. Methods and Data

Belgrade is the largest and most populous city of the Republic of Serbia, situated in Southeastern Europe, with a population of approximately 1.8 million inhabitants. The city encompasses an administrative extent of approximately 3200 km², divided into 17 municipalities. Population density significantly differs among these municipalities. Central districts, including Vračar (approximately 21,000 inhabitants/km²), Zvezdara (around 10,000 inhabitants/km²), and Stari Grad (around 9700 inhabitants/km²), rank among the most densely populated. Conversely, peripheral municipalities like Sopot (approximately 75 inhabitants/km²) and Barajevo (approximately 116 inhabitants/km²) exhibit significantly lower population densities. All municipalities are administratively governed by the City of Belgrade.

The methodology used in this research paper, along with the required datasets, is presented through a methodology workflow (Figure 1). The first step in creating spatial accessibility areas is to acquire the locations for which these areas will be generated. The selected localities included primary healthcare facilities, pharmacies, primary (elementary) and secondary (high) schools, libraries, and green markets (farmers’ markets).

For primary and secondary educational institutions, libraries, and green markets, the addresses of each facility were sourced from the City of Belgrade’s online facility database [40]. In the case of Pharmacy Belgrade, their website [41] provided the addresses of all operational locations. However, the situation for primary healthcare facilities was more complex. Every municipality in Belgrade contains one principal primary healthcare facility, along with numerous smaller, supplementary facilities to serve the entire population within that municipality (the supplementary primary healthcare facilities usually have fewer services than the principal primary healthcare facility). The locations of all facilities (principal or supplementary) were determined individually, with manual searches revealing the locations of supplementary primary healthcare facilities in each municipality or area. The primary objective of the inclusion of primary healthcare facilities was not to conduct a detailed analysis of the services available at each principal and supplementary location as they differ by location; rather, all primary healthcare facilities utilized in this research (principal and supplementary) are categorized under a system that offers “some form” of healthcare without specifically diving into what specific form of healthcare that is. The exclusion of certain primary healthcare facilities occurred when such facilities are situated in industrial zones and cater exclusively to specific industrial companies, thereby not serving the general populace. The exclusion requirements for primary and secondary educational institutions apply to specific facilities, which include vocational schools such as music and ballet academies for gifted pupils, as well as institutions that serve children with disabilities, as that is a topic of our future additional research. In addition, all facilities analyzed are publicly operated, meaning no private entities were considered, as all utilized facilities are accessible to residents at no additional cost.

In the pharmacies category, only Pharmacy Belgrade (ser. Apoteka Beograd) was analyzed. The reasoning behind only analyzing that type of pharmacy is that Pharmacy Belgrade is a public network of pharmacies, which ensures affordable access to medications to the population. The interpretation of the results should not solely emphasize that a specific population lacks access to pharmacies entirely; rather, it should acknowledge that they may not have access to a particular type of pharmacy while still having access to other privately operated pharmacies.

The exclusion criteria for all facilities involved manually verifying the accuracy of the provided information, specifically confirming whether the facility is operational. This was the situation for multiple Pharmacy Belgrade locations, where their website was outdated and listed several facilities that were currently non-operational but had been previously active. This was similarly applied to primary and secondary educational institutions, libraries, and others. Facilities that could not be verified through internet searches or Google Maps were excluded.

The preliminary database, encompassing the name and address of each facility, was created and employed for geocoding purposes, an automated process that generates coordinates from an input address. These coordinates for all facilities were recorded in a preliminary, unverified database. The database verification served two objectives: first, to determine whether the geocoding produced accurate coordinates for the specified facility, and second, to confirm the completeness and accuracy of the City of Belgrade database and other utilized address databases. The verification process required that the location of each facility be cross-referenced with its corresponding location on Google Maps and/or OpenStreetMaps (OSM). If the location was confirmed by either Google Maps or OSM, it was considered verified and suitable for further use. If the specified location was unverified, additional investigation was conducted.

After verifying all locations, the approved database was divided into smaller datasets, each containing the maximum number of data points permissible under the OpenRouteServices Application Programming Interface (API) call limit, as per equation:

$$L\le {\displaystyle \frac{C\times M}{T\times F+B}}$$

(1)

where $L$ represents the maximum number of locations utilized per batch, $C$ denotes the maximum number of API calls permitted per minute, $M$ indicates the duration in minutes of the time window for API calls (established as 1 in this study since batches are dispatched at one-min intervals), $T$ signifies the number of calculated travel times per location, set to 3 in this research, as intervals of 5, 10, and 15 min were employed, $F$ refers to the number of facilities per location, and $B$ is the buffer established to prevent exceeding the limit, set to 1 in this instance. The maximum limit was 40 API calls per minute, necessitating that each batch include no more than 13 locations as per Equation (1), as three API calls are required per location for the three travel-time intervals.

Spatial accessibility areas were generated using ORS isochrone from layer function, which creates an area representing spatial accessibility to specified points within a designated time frame based on a set of parameters. The parameters for this research focused on walking as the mode of transport, with time as the dimension, aiming to determine the area and the corresponding percentage of the population residing within a specified time range from a facility. The intervals were categorized into three groups: 5-, 10-, and 15-min travel-time segments. The last parameter indicated that the specified location would serve as the endpoint, or destination, accessible from the generated area within a designated time interval. The research aimed at determining the percentage of the population residing within 5, 10, and 15 min of a specified facility location.

The post-processing phase was conducted to generate the final spatial accessibility areas. Initially, a layer encompassing all spatial accessibility areas was created through the union of all individual layers. Subsequently, the individual layers were dissolved into a single layer to eliminate potential overlaps. For instance, when two localities are in close proximity, their spatial accessibility areas may overlap, complicating population estimates and potentially leading to an overestimation of the population within the specified area. The final spatial accessibility area was clipped to align with the analysis region, thereby excluding the population residing in less urbanized areas near the research area boundary.

The 2025 Global Human Settlement Layer (GHSL) [42] is a raster dataset with a spatial resolution of 100 m, offering data on the population density within the specified grid. The GHSL was employed to determine the population and, consequently, the percentage of individuals residing within a designated spatial accessibility area. The results were analyzed and interpreted, including analysis of the population outside designated spatial accessibility areas.

All tools employed in this research, including QGIS, ORS, and GHSL, as well as the address databases and supplementary QGIS plug-ins, e.g., Extractor (https://plugins.qgis.org/plugins/extractor/ [43]), are free and open-source software packages. This emphasizes the aim of offering a comprehensive guide, or workflow, for researchers and, more crucially, governmental professionals on conducting spatial accessibility analysis.

Additionally, a survey was conducted where the ranking was employed alongside the 15-min spatial accessibility of facilities from each SAU to assess which SAU possesses the highest overall spatial accessibility. The overall score for each SAU (S) was calculated by:

$$S={\sum }_{i=1}^6(P_i\times R_i)$$

(2)

where $S$ represents the overall score for a specific SAU, $P_i$ denotes the proportion of the population with access to a facility (expressed as a decimal value ranging from 0 to 1), and $R_i$ signifies the ranking coefficient for a particular facility. Equation (2) indicates that the maximum achievable score is 21, provided that the entire population of a specific SAU has access to all six facilities within the stipulated travel-time interval. Conversely, the minimum score is 0 if no members of the population have access to any of the six examined facilities within the 15-min travel-time interval.

3. Results

3.1. Determination of a High-Density Population Cluster in Belgrade

The analysis concentrated only on part of the metropolitan area of Belgrade, rather than the entire region. Figure 2a depicts the entire metropolitan region of Belgrade, overlaid with the GHSL population density raster. As the paper does not encompass the entire metropolitan Belgrade area, a sensitivity analysis was conducted utilizing data derived from the GHSL and official census statistics. The analysis compared the census years of 2000, 2011, and 2022 with the closest corresponding years of the GHSL, namely 2000, 2010, and 2020. The sensitivity analysis was conducted to verify that each reported result is presented with accurate error bounds, specifically regarding the extent and direction of the overestimation or underestimation of the GHSL. The analysis revealed that the GHSL for the three selected periods exhibits an average deviation of 8% in the direction of overestimation. In other words, the results presented in this paper should be interpreted as X-max (8%) of the population that has access to a specific facility, as determined by historical data.

The first phase in defining the analysis region involved establishing the suitable criteria for excluding various municipalities, aiming to ultimately isolate a high-density area within the city of Belgrade. Two criteria were employed for these purposes:

• the population density in all individual areas, whether whole municipalities or SAUs, exceeds 1500 inhabitants per km² according to the Degree of Urbanization classification [44];
• the final area must be contiguous, meaning no exclaves detached from the main high-density population area are permitted.

The initial analysis identified six municipalities that meet the criterion as is (Vračar, Stari Grad, Novi Beograd, Zvezdara, Rakovica, and Savski Venac), while an additional four municipalities require further subdivision into smaller sections, specifically SAUs (Čukarica, Voždovac, Zemun, and Palilula). The remaining municipalities, predominantly suburban, were excluded from the analysis due to being described as large municipalities, ranking from second to ninth in size within Belgrade, yet exhibiting relatively low population density due to a sparse distribution of inhabitants (Figure 2a). The four municipalities requiring additional divisions are predominantly large municipalities in Belgrade, such as Palilula, the largest municipality in Belgrade, encompassing both suburban and densely urban areas that constitute the city center. The same applies to Voždovac, Zemun, and Čukarica. The average population density across the municipality ranges from 1350 to 1215 inhabitants per km², with the exception of Palilula, which decreases to merely 440 inhabitants per km². Consequently, the four municipalities were partitioned into smaller entities, namely SAUs, while densely urbanized areas were retained in the analysis according to the aforementioned criteria, whereas the remainder was excluded (Figure 2b).

Following the initial exclusion of municipalities, the total area of Belgrade decreased by approximately three times, while the average population density throughout the region increased by 2.4 times. The total population expectedly decreased, but by only 22% (Figure 2a,b). Subsequent subdivisions identified a total of 19 units that meet both criteria: a population density exceeding 1500 inhabitants per km² and the formation of a contiguous area of high population density (

Table 1. SAUs and the corresponding areas, number of residents, and population densities; SAU-Sub-administrative unit.

SAUs	Municipality	Area [km2]	Number of Residents [/]	Population Density [Inhabitants/km2]
Vračar	Vračar	2.9	63,153	21,777
Zvezdara	Zvezdara	8.1	86,386	10,665
Stari Grad	Stari Grad	5.3	52,243	9857
Voždovac	Voždovac	11	105,731	9612
Palilula	Palilula	8.8	62,611	7115
Kneževac	Rakovica	6.8	44,005	6471
Novi Beograd	Novi Beograd	40	241,068	6027
Stara Rakovica	Rakovica	6.9	39,227	5685
Čukarica	Čukarica	20	113,407	5670
Mali Mokri Lug	Zvezdara	6.6	31,773	4814
Zemun	Zemun	17	74,140	4361
Jajinci	Voždovac	5.1	21,285	4174
Veliki Mokri Lug	Zvezdara	7	26,519	3788
Mirijevo	Zvezdara	9.2	34,640	3765
Savski Venac	Savski Venac	14	43,163	3083
Resnik	Rakovica	16	43,461	2716
Kumodraž	Voždovac	10	26,179	2618
Železnik	Čukarica	15	34,629	2309
Višnjica	Palilula	11	24,466	2224

, Figure 2c). The final exclusion of non-high-density regions in Belgrade resulted in a reduction of the total area by approximately 14 times, while the population decreased by only 39%, from 1.8 million to 1.1 million residents. The chosen area represents a very high-density zone, specifically the urban zone of Belgrade, with an area of 225 km². The research area, being 14 times smaller than the entire metropolitan area of Belgrade, while still containing the majority of the population, confirms its suitability for spatial accessibility analysis.

3.2. Spatial Accessibility of Primary Healthcare Facilities and Pharmacies

Figure 3 illustrates the spatial accessibility areas for travel-time intervals of 5, 10, and 15 min to primary healthcare facilities and pharmacies. The total number of primary healthcare facilities in the designated research area equated to 57. With the 5-min travel-time interval, the total number of inhabitants that could reach a primary healthcare facility was 194,000, representing approximately 16.6% of the overall population within the study area. The area covered with the 5-min travel-time interval constituted 7.3% of the entire research area or approximately 16.4 km². The 10-min travel-time interval resulted in a figure of 577,000 individuals, representing approximately 49.3% of the population of the research area, and the area covered was approximately 24.8%, equivalent to around 55.9 km². Extending the travel-time interval from 5 min to 10 min, thereby doubling the interval, resulted in reaching approximately 382,000 additional individuals, representing a 32% increase in coverage. The doubling of the travel-time interval also led to a 3.4-times increase in the area covered. The largest travel-time interval, specifically the 15-min interval, encompassed approximately 922,000 individuals, representing 78.9% of the total population within the research area. Approximately 46% of the research area, equivalent to 105 km², was covered with the largest travel-time interval.

Approximately 246,000 individuals, representing 20.1% of the population, are unaccounted for within the largest 15-min travel-time interval. The SAU of Železnik exemplifies this situation, as it contains only one primary healthcare facility serving approximately 17,000 individuals out of a total population of 34,000. Consequently, only 50% of Železnik’s residents have 15-min access to primary healthcare services. The SAU of Kneževac and Resnik display similar spatial accessibility to Železnik, as these suburban areas possess only one primary healthcare facility serving the entire population.

Conversely, regions such as Vračar, Zvezdara, Stari Grad, Voždovac, and Palilula exhibit higher inhabitant numbers and thus possess a larger amount of primary healthcare facilities, resulting in full areal coverage. Additionally, areas such as Vračar, Zvezdara, Stari Grad, Voždovac, and Palilula display the effect that the entire population is within a 15-min distance from some primary healthcare facility, even if it is not within the same SAU, i.e., primary healthcare facilities’ spatial accessibility areas overlap in those regions.

Pharmacies display a comparable number of facilities to primary healthcare facilities, with the total count of pharmacies amounting to 55. In comparison to primary healthcare facilities, pharmacies exhibit marginally greater spatial accessibility within the 5-min travel-time interval, at 20.28%, which corresponds to approximately 237,000 individuals. The difference in spatial accessibility between pharmacies and primary healthcare facilities is minimal for the 10-min travel-time interval, with pharmacies at approximately 50.23% (587,000 individuals) and primary healthcare facilities at 49.3% (577,000 individuals). Within the 15-min travel-time interval, primary healthcare facilities exhibit better spatial accessibility, demonstrating a difference of approximately 5%, with pharmacies at 73.77% compared to 78.94% for primary healthcare facilities. The area encompassed within a 5-min travel time is approximately 7.3% of the total area for both pharmacies and primary health care facilities. A comparable scenario arises for the 10-min travel-time interval (22.55%). For the 15-min travel-time interval, primary healthcare facilities encompass approximately 4.8% more area than pharmacies.

Approximately 25% of the population, or roughly 306,000 individuals, are unaccounted for within a 15-min travel time to the nearest pharmacy. For SAUs like Železnik and Resnik, the situations remain unchanged compared to those related to the primary health center example. The circumstances in Kumodraž are noteworthy, as the SAU contains three health centers, Kumodraž 1, 2, and 3, yet only Kumodraž 1 houses a pharmacy within the primary health center. The SAU of Jajinci has one primary health center that serves most of its population within a 15-min travel-time interval; however, it lacks a pharmacy, resulting in the community’s lack of access to pharmaceutical services. Moreover, significant areas of Novi Beograd, Savski Venac, and the entirety of Višnjica are devoid of spatial coverage by pharmacies.

3.3. Spatial Accessibility to Primary and Secondary Educational Facilities

Elementary schools constitute the largest category of facilities examined, totaling 101 within the research area. Therefore, they exhibit the highest spatial accessibility across all travel-time intervals among the facilities examined. At a travel-time interval of 5 min, the overall percentage of population accessibility is 31.2%, representing an increase of over 11% compared to primary healthcare facilities and pharmacies. During the 10-min travel-time interval, the population served is 71.4%, approximately 835,000 individuals, which is comparable to the 15-min travel-time interval for primary healthcare facilities and pharmacies. The largest travel-time interval encompasses approximately 1 million individuals who can access a primary elementary school within 15 min, representing 89.5% of the total population in the research area. Primary schools occupy the broadest area within each travel-time interval, with percentages of 12.65% (28.5 km²), 35.5% (80 km²), and 56.4% (126.2 km²) for the 5-, 10-, and 15-min intervals, respectively (Figure 4).

However, although elementary schools serve 89.5% of the population within the largest travel-time region, approximately 122,000 individuals remain unaccounted for. Every SAU contains a minimum of one elementary school (Resnik, Jajinci, Kumodraž, Mali Mokri Lug, and Veliki Mokri Lug with one), while municipalities such as Novi Beograd, Vračar, and Zvezdara exhibit a greater number of elementary schools. The inconsistency in population coverage is evident in the aforementioned SAUs, which possess only one elementary school, resulting in unserved population pockets that are located around the 15-min travel-time area. A good example of the disparity is evident in the SAU of Resnik, where just one elementary school serves a population of 23,000 individuals within a 15-min travel-time radius, constituting only around 50% of the total population of 46,000 in Resnik.

High schools constitute one of the larger categories of facilities among all groups examined in this research paper, with a total number of 55 facilities. Despite having a relatively high total number of facilities in the analysis region, the percentage covered within the specified travel-time intervals reveals significant discrepancies when compared to other facilities with fewer total establishments. For instance, the proportion of the population with 5-min spatial accessibility to a high school is approximately 15.56%, which is lower than that of pharmacies and comparable to primary healthcare facilities. As the travel-time interval increases, the discrepancy also increases; for instance, a 10-min travel-time interval corresponds to 36.15% of the population covered, whereas the coverage for primary healthcare facilities and pharmacies is approximately 13% higher within the same interval. The greatest discrepancy occurs during the 15-min travel-time interval, with differences of approximately 21% and 26% when compared to pharmacies and primary healthcare facilities, respectively. The 15-min travel-time interval encompasses approximately 52.5% of the total population within the analysis region, amounting to roughly 613,000 individuals.

The substantial discrepancy between high schools and other facilities, despite high schools possessing a comparable number of facilities in comparison to other facility groups, is illustrated in Figure 4 when analyzing the distribution of high schools within the research area. Initially, the SAUs of Višnjica, Mali Mokri Lug, Veliki Mokri Lug, Kumodraž, Jajinci, and Resnik lack high schools within their respective boundaries. These municipalities comprise a total of 174,000 individuals lacking walking access to a high school. The spatial distribution of high schools is concentrated within six SAUs (Stari Grad, Palilula, Vračar, Savski Venac, Voždovac, and Zvezdara), housing a total of 35 high schools. These six SAUs comprise 35% of the total population of the analysis region yet encompass approximately 64% of all high schools. In comparison, two of the largest SAUs by population size, Novi Beograd and Čukarica, have only seven and two high schools, respectively. The spatial distribution was the exclusive factor contributing to the limited spatial accessibility of high schools in the analyzed region and the significant disparities observed in comparison to other facilities.

This can be expanded on by examining the percentage of the 15–19 age demographic (typically high school age in Serbia) across municipalities [45]. Novi Beograd was the largest municipality in terms of the percentage of the 15–19 age group, comprising 0.54% of the total population of metropolitan Belgrade, which equates to approximately 9001 individuals aged 15–19. The second entry on the list was Palilula, comprising 0.53% or 8914 individuals aged 15–19; however, it is important to note that this figure pertains to the entire Palilula municipality, while this paper only addresses the urban core of Palilula, thus excluding it from this discussion. In contrast, municipalities with larger number of high schools, namely Stari Grad (7) and Savski Venac (6), exhibit the lowest percentages of individuals aged 15–19, specifically 0.1% (1714) and 0.1% (1685), respectively, out of the entire population of metropolitan Belgrade. Novi Beograd exhibits the highest percentage of high school-aged individuals at 0.54% and ranks among the areas with the largest number of high schools, alongside Stari Grad, Zvezdara, and Voždovac, each having seven. Conversely, Čukarica possesses only two high schools and ranks fourth in terms of the percentage of high school-aged individuals (third, excluding Palilul), accounting for 8366 or 0.5% of the total metropolitan Belgrade population. This highlights the significance of better high school development in the municipality to align with comparable municipalities, such as Zemun (six high schools) and Zvezdara (seven high schools), which have similar percentages of high school-aged populations, specifically 0.51% (8543) and 0.48% (8008), respectively.

In spatial terms, the accessibility areas of high schools occupy between 6% (13.46 km²) and 16.7% (37.6 km²) to 28.79% (64.82 km²) of the analyzed region’s area.

3.4. Spatial Accessibility to Libraries and Green Markets

Figure 5 illustrates the spatial accessibility zones to libraries and green markets, where it can be seen that the research area contains a total of 41 libraries. The research area contains 14 fewer libraries than high schools (55 high schools compared to 41 libraries); however, libraries exhibit similar spatial accessibility to high schools, and in the largest travel-time interval, they demonstrate better spatial accessibility. For the 5-min travel-time interval, approximately 15.3% of the total population is covered, while for the 10-min travel-time interval, this figure rises to around 40%. Both percentages are relatively comparable to those exhibited by the accessibility areas of high schools. During the 15-min travel-time interval, approximately 707,000 individuals, or 60.5% of the population, are encompassed, indicating better population coverage compared to high schools by roughly 8.1%. The area coverage values closely resemble those of libraries: 5.6% for the 5-min travel-time interval, 17.8% for the 10-min travel-time interval, and 32.6% for the 15-min travel-time interval.

Like the previous examples, SAUs such as Kneževac, Resnik, Jajinci, Kumodraž, Veliki Mokri Lug, Mali Mokri Lug, and Višnjica lack libraries within their boundaries, whereas Vračar, Zvezdara, and Stari Grad possess multiple facilities.

Green markets were the final facility examined in this research paper and represented the facility type with the fewest establishments in the research area, totaling only 27. The 5-min travel-time interval indicated a population coverage of 9.5%, signifying that the green market facility was the sole facility with accessibility below 10% among all presented facilities. The 10-min travel-time interval exhibited a significant increase compared to the 5-min travel-time interval, where the population covered was 33.8%, a figure that is merely 2.4% lower than the spatial accessibility of high schools for the same travel-time interval. The 15-min travel-time interval indicated a population coverage of 62%, approximately 724,000 individuals. This figure denotes a 9.5% increase compared to the same travel-time interval for high schools and is analogous to the spatial accessibility of libraries. The covered area percentages were 3.88% (8.73 km²), 15% (33.7 km²), and 32.7% (73.5 km²) for the 5-, 10-, and 15-min travel-time intervals, respectively. Similarly to the prior examples, the SAUs of Resnik, Jajinci, Kumodraž, and Višnjica lack green markets within their respective areas.

3.5. Survey-Informed Analysis of Individual Region Inequalities in Spatial Accessibility

A survey was administered to 31 participants who were voluntarily asked to participate in the survey to evaluate the significance of having each facility situated near their residence. The predominant respondents were aged 18–25 (13) and 26–35 (10), while the remaining participants were allocated among the 36–45, 46–55, 56–65, and 65+ age groups (8). Approximately 58% (18) of the respondents were male, while the remaining 42% were female (13). The educational distribution included 22 individuals with college or university education, five with postgraduate degrees, and four with secondary education.

The survey required participants to rank facilities by importance, with the most significant facility, based on proximity to their residence, assigned a grade of 1, while the least significant facility received a value of 6. The facility with the lowest total score is deemed the most essential to be in proximity to the participants’ residences. The primary healthcare service facility was deemed the most significant, with 15 out of 31 participants ranking it in the top position, i.e., most important to the survey participants to be within 15-min walking distance to their home. The overall score of primary healthcare service facilities was 70. Elementary schools rank second with 89, while pharmacies occupy third place with 91, exhibiting a relatively narrow margin between the two. The remaining locations were occupied by green markets (101), high schools (135), and, lastly, libraries (165).

Three SAUs exhibit superior results, each achieving over 90% population coverage across all six facility categories (

Table 2. Population coverage per SAU for the 15-min travel-time interval and calculated score per SAU; SAU-Sub-administrative unit; P25, P50, and P75 correspond to the 25th, 50th, and 75th percentile, respectively.

Cadaster Municipality	Municipality	Primary Healthcare Facility	Elementary School	Pharmacy	Green Market	High School	Library	Score
Vračar	Vračar	100.00	100.00	100.00	99.81	100.00	100.00	20.99
Stari grad	Stari grad	99.24	98.88	98.43	93.03	98.84	99.20	20.60
Zvezdara	Zvezdara	99.80	100.00	91.04	90.38	97.92	96.55	20.26
Voždovac	Voždovac	98.00	97.99	91.36	90.64	55.10	78.29	19.04
Mirijevo	Zvezdara	92.31	96.14	66.30	72.36	71.41	79.37	17.43
Zemun	Zemun	81.44	98.36	85.27	53.31	69.29	80.47	17.39
Palilula	Palilula	85.99	97.43	82.30	73.62	68.64	52.73	17.00
Savski venac	Savski venac	91.35	91.12	52.03	70.00	78.19	69.62	16.48
Stara Rakovica	Rakovica	88.89	96.44	60.48	45.91	75.02	49.45	15.95
Čukarica	Čukarica	68.67	91.44	81.61	62.22	30.00	74.13	15.16
Novi Beograd	Novi Beograd	69.98	90.45	69.45	47.77	49.21	57.49	14.49
Kneževac	Rakovica	52.28	84.26	86.23	78.24	41.20	9.30	14.06
Železnik	Čukarica	50.26	56.48	48.58	57.72	21.81	53.82	12.88
Mali Mokri Lug	Zvezdara	56.27	76.46	61.00	60.44	6.19	22.87	11.80
Veliki Mokri Lug	Zvezdara	88.08	76.08	56.66	50.14	0.00	2.12	10.76
Kumodraž	Voždovac	82.32	60.69	62.53	6.39	0.00	9.11	10.49
Višnjica	Palilula	55.63	84.76	7.90	20.60	3.22	2.64	8.60
Jajinci	Voždovac	71.12	61.65	10.90	2.07	0.00	0.25	7.85
Resnik	Rakovica	42.01	54.25	42.01	0.00	1.07	0.72	6.94
		Min	P25	P50	P75	Max

). Vračar stands out as the foremost, nearing 100% in all categories, followed by Stari Grad and Zvezdara. Vračar achieving the highest overall score was expected given that it is the smallest SAU by area; conversely, the three SAUs with the best scores also exhibit the highest population densities (Table 1). Voždovac, the fourth-ranked SAU, achieved high scores for the four most significant facilities, each with a population coverage of 90%. However, it performed relatively poorly in terms of high school coverage at 55% and library coverage at approximately 78%. In contrast, Resnik received zero scores for facilities such as green markets, high schools, and libraries while achieving approximately 40–50% for primary healthcare facilities, elementary schools, and pharmacies.

Figure 6a illustrates the spatial distribution of the computed score within the research area of Belgrade. The four highest-performing SAUs are clustered together, surrounded by relatively well-performing SAUs based on the overall score. In contrast, from the southwest to the northeast on the edge of the research area, there are poorly performing SAUs, with the exception of Mirijevo, which ranks as the fifth-best overall, following Voždovac.

The survey facilitated the categorization and subsequent analysis of regions within the study area where multiple facilities are unreachable within a 15-min travel time. Figure 6b illustrates the regions corresponding to the combining of primary healthcare facilities and elementary schools. The region depicted in Figure 6b comprises approximately 90,000 individuals, with around 52% of this population residing in Novi Beograd (18,000), Resnik (17,000), and Železnik (11,000). Figure 6c illustrates identical information for the combination of pharmacies and green markets, covering an area of approximately 240,000 individuals, distributed as follows: Novi Beograd (64,000), Resnik (25,000), and Višnjica (19,000), among others. Figure 6d illustrates the identical no-service area for high schools and libraries, which were ranked as the two lowest facilities in the survey. The total population of the area is 358,000, with the majority residing in Novi Beograd (52,000), Resnik (43,000), and Čukarica (29,000). Figure 6e illustrates the worst-case scenario, specifically the regions within the analysis area where all six facilities exceed the 15-min travel time. The total population in those areas is 78,000, predominantly situated in Resnik (17,000), Novi Beograd (14,000), and Železnik (11,000).

To improve the understanding of spatial equity, we compared our accessibility findings with the average income levels across Belgrade’s municipalities [46]. The most substantial salaries are found in central municipalities, including Stari Grad (RSD 154,469), Vračar (RSD 150,205), Novi Beograd (RSD 142,355), and Savski Venac (RSD 140,472). Conversely, peripheral municipalities like Rakovica (RSD 96,805), Palilula (RSD 97,578), and Zemun (RSD 101,277) exhibit lower average incomes. Notably, these lower-income municipalities tend to overlap with regions of poorer access. This overlap signifies that spatial accessibility deficiencies are frequently experienced disproportionately by inhabitants of municipalities with lower average incomes, emphasizing the necessity of incorporating spatial justice in urban development. It is important to note that for municipalities where only the urban segment was evaluated, such as Palilula and Voždovac, the reported salaries pertain to the entire municipality rather than solely the urban segment. These examples should be considered; however, official data are presented in aggregate form rather than at a more granular spatial resolution. Consequently, the outcomes for these municipalities should be regarded with caution. It should be noted that not all patterns align perfectly; for instance, Novi Beograd represents an outlier where the study design may insufficiently capture the socioeconomic–spatial relationship, a subject beyond the scope of this paper but requiring future, additional research.

4. Discussion, Limitations, and Future Prospectives

The importance of these results presented in this research paper cannot be overstated, as they offer a precise method for analyzing and conveying areas that are inadequately represented in terms of walking accessibility to a specific facility or an array of facilities. The results distinctly indicate SAUs where nearly the entire population resides within a 15-min travel time to all facilities, including Zvezdara, Vračar, and Stari Grad. Conversely, the results distinctly indicate which SAUs exhibit inadequate spatial accessibility to specific facilities, including Resnik, Jajinci, and Višnjica. These results are highly valuable for decision-makers as they facilitate data-driven decision-making, e.g., for the location problem of healthcare centers [47] or other facilities. Conversely, the results also indicate the requisite advancements in mathematically driven modeling, as the foundational methods for generating spatial accessibility areas are based on mathematical techniques. The advancement of these methods to include daily traffic fluctuations and other factors that directly affect accessibility zones may serve as a promising area for research exploration.

Regarding the concept of a 15-minute city, three municipalities, Vračar, Zvezdara, and Stari Grad, are identified as nearly fully integrated into this framework for the evaluated facilities. The three SAUs indicate that for the six examined facilities, they conform to the 15-minute city concept; however, further research is required to assess additional essential facilities for the overall well-being and quality of life of residents, including grocery stores, kindergartens, green spaces, parks, public transport stops, and others. The notion of a 15-minute city is a crucial research area and is essential for policymakers, as a walkable urban environment mitigates traffic congestion, encourages active and healthy living, and can alleviate issues such as air pollution [48]. Spatial accessibility can be a vital tool in realizing the objective of a walkable, 15-min area.

This research aims to present the methodology section as a workflow, as it is believed that this approach would be most advantageous for other researchers, and particularly for governmental professionals seeking a step-by-step guide on conducting spatial accessibility analysis. Spatial statistical methods are essential for data-driven decision-making and should be integral to policy formulation. Consequently, the methodology was presented as a workflow designed to be inclusive for a diverse range of researchers and governmental employees, with the objective of encouraging a broad spectrum of individuals to integrate these methods into their analyses. Moreover, it is important to acknowledge that the foundational mathematical principles governing accessibility area generation can be complex and elusive for most individuals outside the relevant field. However, the application of these methods is comparatively easy using tools like QGIS and ORS. Therefore, it is advantageous to present the methodology as a workflow to enhance inclusivity. This research exclusively employed free and open-source software, including QGIS for data visualization and analysis, ORS for spatial accessibility analysis, and other freely available tools to achieve inclusivity.

Moreover, spatial accessibility analysis serves as a crucial instrument for evaluating advancements or identifying deficiencies in the attainment of various SDGs. The SDGs encompass SDG 1.4, which mandates that all individuals have access to basic services, allowing for the assessment of the population lacking such access. Similarly, SDG 3.8 can be evaluated via spatial accessibility analysis, specifically regarding the coverage of essential health services, as demonstrated in this research through the examination of primary healthcare facilities and pharmacies. Likewise, SDG 11.2 can be examined through the perspective of spatial accessibility, as its objective is to establish safe, affordable, accessible, and sustainable transportation systems for all. A comparable scenario exists for SDG 11.7, which pertains to the accessibility of green and public spaces. All previously mentioned factors and additional elements can be carefully monitored through spatial accessibility, particularly as tools like ORS and QGIS facilitate this process in a cost-effective and straightforward manner.

This research examined spatial accessibility to six facilities; however, additional analysis is necessary, including the spatial accessibility of kindergartens, banks, post offices, grocery stores, emergency services, special needs educational facilities, and others, which is vital for the overall well-being of the population in a specific area. This paper exclusively analyzed the strict urban zone of Belgrade; additional research is required to encompass the entire metropolitan area of Belgrade, as well as to examine other urban regions in Serbia, including two other large cities in Serbia, such as Novi Sad and Niš, and to assess rural areas in the Republic of Serbia regarding spatial accessibility to essential services.

The results provide policymakers with a baseline reference, highlighting areas lacking access to specific facilities or groups of facilities. The findings indicated that 246,000 individuals lack access to primary healthcare facilities, while 122,000 individuals lack access to elementary schools near their homes. Underrepresented areas were identified, which can facilitate the development of strategies to enhance accessibility, informed by the spatial distribution of these populations. The same applies to the other facilities discussed in this paper, as well as to additional facilities for which spatial accessibility analysis can be conducted.

It is interesting to note that there have been no similar studies for Serbia and Belgrade with the level of detail and comprehensiveness presented in this paper. There are several studies that have addressed the issue of accessibility for Serbia/Belgrade, in which only health centers were analyzed. For example, Valjarević et al. [49] analyzed the accessibility of health centers in the city of Belgrade and showed that it takes an average of 100 min to reach a health facility on foot with the best-served central municipalities of Belgrade (Stari Grad, Vračar, Savski Venac, etc.). These results are in line with our research when we compare urban with suburban areas, especially if we take into account the fact that a longer average time is needed in suburban areas of the city (Resnik, Železniki, Kneževac, Višnjica, etc), which can explain the slightly longer average walking time (100 min) to a health facility from Valjarević et al.

Another study that addressed the same issue analyzed accessibility across the entire territory of Serbia, for intervals of 10 min to an hour, when traveling by car [50]. The research showed that the best-served parts of the country are the central areas of the largest cities in Serbia, where accessibility is measured in time intervals of less than 10 min. Although the studies are not formally the same, a parallel can be drawn indicating that central city areas have significantly better accessibility to healthcare institutions, whether on foot or using some means of transport (car or public transport).

In addition to the aforementioned analysis, a significant facet of spatial accessibility analysis, as well as spatial analysis in general, and a topic for future research, pertains to the spatial accessibility of betting locations in Belgrade. The three highest-ranked SAUs exhibit the following number of physical gambling facilities: 48 in Vračar, 40 in Stari Grad, and 89 in Zvezdara [51]. Spatial accessibility analysis can be employed to assess service overlap in the locations of gambling establishments, specifically regarding the overrepresented population (those most likely to encounter gambling venues) and, most importantly, the spatial accessibility of gambling facilities in relation to elementary and secondary schools. This research is crucial for both the general populace and policymakers to protect individuals from gambling addiction, since proximity to physical gambling locations can be associated with a higher likelihood of gambling [52,53,54].

Apart from the direct findings of this study, it is also important to situate spatial accessibility analysis within the wider spectrum of transport and urban planning research. For instance, recent investigations on public transit design in Greater Cairo [55], city of Makarska [56], Suzhou [57] have emphasized the role of integrated underground and bus systems, or even “ride-pooling” in reducing passenger transfers and enhancing network connectivity. Also, the integration of walking and bicycling facilities with public transport stations is especially important to maintain functionality of the transport system on the one hand, and to provide high-quality service to users on the other [58,59]. Similar methodological approaches that focus on minimizing transfers in large-scale subway networks demonstrate how mathematical formulations and strategic design can significantly increase system efficiency and user comfort [60]. While these studies primarily address mass transit optimization rather than neighborhood-level accessibility, they illustrate the importance of holistic planning frameworks that account for both efficiency and equity in urban mobility. In parallel, emerging research on travel demand forecasting highlights the growing incorporation of machine learning [61] and microsimulation tools [62,63] to improve prediction accuracy in multimodal networks [64]. Although such techniques were not applied in the present analysis, they underscore the ongoing methodological advances in transport planning that could, in the future, be coupled with spatial accessibility studies to provide richer insights into demand dynamics and infrastructure needs.

The paper’s limitations must also be acknowledged. The principal limitation of the paper lies in the quality of the input data, specifically the repository of facility locations in the city of Belgrade. Multiple methods were employed to ensure comprehensive inclusion of all locations in this analysis, including cross-referencing data from the city of Belgrade with Google Maps and OpenStreetMap; however, there exists a possibility that certain locations were neglected or that some facilities that should have been excluded were not. The quality of input data for spatial accessibility analysis is crucial, as the generation of spatial accessibility areas depends on the input locations provided to the model. As outlined in the methodology and reiterated in the discussion, cross-referencing the input data with other services is regarded as best practice.

In addition to the constraints of the input data, it is important to acknowledge that the population distribution raster from the GHSL is not entirely exact in reflecting the total population of a specific area when compared with other sources, including official governmental data. This bias can only be addressed with an improved population distribution dataset; in its absence, the GSHL remains fairly precise, readily accessible, and easily incorporated into QGIS workflows. Furthermore, it is important to acknowledge that the GHSL depicts population distributions at a 100 m resolution, and higher-quality datasets on population distribution are scarce and infrequent for extensive areas such as the urban city core of Belgrade or larger.

A further limitation of the study is that the examined facilities are state-operated, as schools in the Republic of Serbia are predominantly state-run, with a limited number of privately operated institutions. The analyzed pharmacies were exclusively state-run, while private pharmacies were not included in the analysis. The research findings should not be interpreted to imply that residents of a particular excluded area lack access to all pharmacies; rather, they lack access to a particular type of pharmacy. Analogously, green markets exemplify a distinct market category, and residents of a particular excluded area likely possess access to some form of grocery store as an alternative.

The limitations of OpenRouteService, the QGIS plug-in employed for generating accessibility areas, must be recognized. The OpenRouteService plug-in for walking mode employs a constant walking speed of 5 km/h in all instances, which may be regarded as a limitation. Research on walking speeds, categorized by age and gender, indicate that the average walking speed for males varies from 4.76 km/h (in their 70s) to 5.26 km/h (in their 40s), whereas for females, it ranges from 4.58 km/h (in their 70s) to 5.09 km/h (in their 30s) [65]. A subsequent meta-analysis of walking speed indicated ranges of 3.48 km/h for males aged 80 and older to 5.16 km/h for males aged 30–59; for females, the corresponding values were 3.39 km/h (aged 80 and older) to 5.0 km/h (in their 40s) [66]. This research indicates that the expected mean percentage error in walking speed is approximately 2.9% for males and 3.3% for females, [65], or about 9.97% for males and 13.71% for females [66]. A significant factor to note is that the most pronounced discrepancies are typically found among individuals aged 60 and above, who comprise roughly 26% of the total Belgrade metropolitan population, whereas those under 19 years of age account for an additional 20% [45]. Consequently, for a slight majority of approximately 54% of the population, the divergence between the modelled walking speed and the true velocities documented by Bohannon [65] and Bohannon and Andrews [66] is negligible, approximately 2.5%. This can be further displayed using the time values presented in this paper. The most significant observed discrepancy was approximately a 7-min difference for the female age group of 80 and older. For a modeled 15-min walking, the actual walking time for an 80-year-old female would be approximately 22 min, indicating a 7-min (or 47%) discrepancy. It is important to note that the population aged 80 and above constitutes only 4.27% of metropolitan Belgrade, with females in this age bracket representing only 2.7% [45], indicating that this demographic, despite exhibiting the most significant deviation, comprises a minor segment of the overall population. Other discrepancies presented as percentage errors can likewise be represented in time values. Males in their 20s would reach the modeled 15-min accessibility area threshold in 15 min and 20 s, while those in their 30s, 40s, and 50s would do so in 14 min and 32 s [66]. Likewise, women in their 20s and 30s would reach the same 15-min accessibility area threshold in 15 min 31 s and 15 min 35 s, respectively. Smaller discrepancies were noted in the Bohannon [65] data. This illustrates that, although some discrepancies arise in modeling accessibility areas, they are generally minor in most instances. Furthermore, neither Bohannon [65] nor Bohannon and Andrews [66] displayed the impact of weather conditions on variations in walking speed, which can negatively affect walking speeds [67]. Willberg et al. [68] conducted research demonstrating variations in walking speeds among adults and individuals over 65 years old under dry winter conditions, compact snow, 2–8 cm of snow, and gritted ice conditions. The most significant disparity for adults was from 5.15 km/h under dry conditions to 4.48 km/h on winter gritted ice, representing a 13% difference. The discrepancy for the population aged over 65 years was more pronounced, ranging from 4.26 km/h in dry conditions to 3.5 km/h on winter gritted ice, representing an 18% difference. The prior analysis indicates that the elderly exhibit the most significant discrepancy, a finding supported by research conducted by Fitzpatrick et al. [69], which reveals that the elderly or physically disabled individuals have slower walking speeds, with a median value of 3.71 km/h, signifying that 50% of this population walks at speeds below this value. The studies by Bohannon [65] and Bohannon and Andrews [66] did not present data for children under 12 years and teens aged 12–18. In contrast, Fitzpatrick et al. [69] reported a median walking speed of 4.8 km/h for children and approximately 5.09 km/h for teens. Walking speed exhibits considerable variability, influenced by factors such as age, sex, and weather conditions.

To achieve the most precise modeling outcomes, pedestrians should be considered diverse rather than uniform in their walking speed [70]. This paper aimed to model accessibility areas for different facility types within a large 225 km² region of Belgrade’s urban center. Modeling extensive areas necessitates some degree of simplification, including the exclusion of finer details like variations in walking speed. The walking speed of the majority of the population of Belgrade’s metropolitan area aligns with the standard 5 km/h value established by OpenRouteService, leading to negligible errors. For a minority of the population, significant discrepancies may arise, exemplified by the largest discrepancy of 7 min (47%) observed in females aged 80 and older. Nonetheless, when employing modelling tools to extensive areas, a certain level of discrepancies and approximations is expected. It could be recommended for OpenRouteService to incorporate an option enabling users to specify custom walking speeds. Utilizing the data from Bohannon [65] and Bohannon and Andrews [66] and other sources previously displayed would facilitate more accurate modeling, allowing for the precise representation of various age groups, regardless of their size within a population. In the absence of such functionality, the standard walking speed of 5 km/h is fairly precise for the majority of the population and for an approximation of fair-weather conditions. Future research will concentrate on providing a thorough analysis of discrepancies among age groups, walking speeds, sexes, and other variables within the default walking speed of 5 km/h in OpenRouteServices, especially from data obtained in Serbia, as regional differences can play a role in walking speed [67]. However, this brief discussion exclusively utilized the 15-min travel time as an example showing the largest errors between modeled and true walking speeds.

The future potential of the included survey can be enhanced in both scale and inquiry to more accurately assess the population’s inclination towards the presence of a specific type of facility near their residence. Inquiries regarding whether respondents have children, whether their preference for a particular type of facility location has evolved with age, and the inclusion of additional facilities in the questions can be integrated into future research and may constitute independent studies. To assess the overall ranking of the facilities, the quality and quantity of the respondents are adequate.

Spatial accessibility is a significant area of research, applied by various researchers and practitioners in their work. Subsequent research will aim to broaden the research scope and integrate additional facilities in the future.

5. Conclusions

This study represents the first comprehensive spatial accessibility analysis of key public facilities in Belgrade, Republic of Serbia. Focusing on primary healthcare centers, pharmacies (Pharmacy Belgrade), primary and secondary educational institutions, libraries, and green markets, the research revealed critical insights into the distribution and accessibility of these essential services across the urban area.

Central municipalities such as Vračar, Stari Grad, and Zvezdara largely conform to the 15-minute city concept, where almost the entire population can access all analyzed facilities within a 15-min walking distance. In contrast, peripheral areas like Resnik, Jajinci, and Višnjica showed clear gaps in service accessibility, highlighting underrepresented regions that require targeted policy attention and investments to improve residents’ welfare.

A major strength of this paper lies in its direct and straightforward methodological approach, specifically designed to be easily understandable for a broad audience, especially decision-makers. The entire workflow relies on free and open-source software, datasets, and tools, demonstrating that comprehensive spatial analyses can be performed without financial or technical constraints. This approach ensures that governmental professionals, researchers, and community organizations can replicate and apply this method to other contexts, encouraging wider adoption of spatial planning tools.

The key contributions of this research include:

• A comprehensive spatial analysis of multiple public facilities, identifying service coverage and gaps within Belgrade’s urban area.
• Direct policy recommendations aimed at improving spatial equity, especially in underserved peripheral areas.
• A replicable, open-access workflow that empowers diverse stakeholders to conduct similar analyses using free tools and data.
• Promotion of data-driven urban planning research with practical applications for the benefit of the wider population.

This study lays a strong foundation for future research in Belgrade and Serbia, enabling further exploration of spatial accessibility across various cities and rural regions, and fostering comparative analyses between urban and rural populations. Through its accessible and open nature, this paper contributes meaningfully to advancing spatial justice and sustainable urban development.