Irregular Area Cartograms for Local-Level Presentation of Selected SDGs Indicators Based on Earth Observation Data

Abstract

The objective of this study is to explore the applicability of irregular area cartograms for the visualization of sustainable development indicator components, utilizing earth observation (EO) data. The analysis focuses on selected Sustainable Development Goals (SDG 11 ‘Make cities and human settlements inclusive, safe, resilient and sustainable’ and SDG 13 ‘Take urgent action to combat climate change and its impacts’) and specific targets and indicators related to green urban areas and air quality (targets: 13.2, 11.6, and 11.7; indicators: 11.6.2., 11.7.1., 13.2.2.). A comprehensive review of the relevant literature indicates that irregular area cartograms are employed only sporadically in the context of SDG monitoring, particularly at lower levels of territorial division (i.e., communes and counties). To address this gap, a series of thematic maps, including choropleth maps and irregular area cartograms, was developed. These visualizations are based on EO-derived datasets and supplemented with statistical information obtained from the Local Data Bank of the Statistics Poland. The analysis demonstrates that irregular area cartograms provide an effective means of visualizing spatial disparities in variables such as urban green space availability and air pollution at the commune and county levels. These visualizations enhance the interpretability of complex indicator structures and support more nuanced assessments of progress toward selected Sustainable Development Goals, especially in spatially detailed analytical frameworks. Preliminary usability testing among potential users revealed that irregular area cartograms are perceived as an interesting visualization technique that enhances data interpretation.

1. Introduction

Monitoring the achievement of the Sustainable Development Goals (SDGs) is an important topic that can be assisted by using indicators elaborated based on Earth Observation (EO) data. The United Nations (UN) Sustainable Development Goals are global development targets that were adopted in 2015. All countries have agreed to work towards achieving them by 2030 [1]. The Sustainable Development Goals (SDGs) form a closely linked set of priorities, where progress in one area often affects outcomes in others. Reaching these goals requires finding a practical balance between environmental protection, economic growth, and social well-being. Their implementation depends not only on political will but also on the involvement of institutions, local governments, and civil society. To monitor progress, the UN Secretary-General releases an annual report prepared in collaboration with the UN system, using data collected by national statistics authorities and regional partners [2]. These reports contain information on the completion of particular SDGs, targets, and indicators, part of which is presented in maps. Statistics Poland (SP) is responsible for presenting the implementation of the SDGs in Poland. On the SP’s website, information on 250 global indicators can be found, including data for Poland regarding monitoring the Sustainable Development Goals.

The study presents indicators related to SDG11 (Goal 11—Sustainable cities and communities) and SDG13 (Goal 13—Climate action). As a preliminary step, the following targets and indicators were selected for analysis:

• Target 11.6: Reduce the environmental impacts of cities (Indicator 11.6.2—Annual mean levels of fine particulate matter);
• Target 11.7: Provide access to safe and inclusive green and public spaces (Indicator 11.7.1—Average share of the built-up area of cities that is open space for public use for all);
• Target 13.2: Integrate climate change measures into national policies, strategies, and planning (Indicator 13.2.2—Total greenhouse gas emissions per year).

Air pollution remains a major challenge in urban areas. Elevated concentrations of particulate matter (PM), nitrogen oxides (NO₂), tropospheric ozone (O₃), and sulphur dioxide (SO₂) are known to negatively impact human health and quality of life in affected regions. These pollutants also contribute to the degradation of biodiversity and the disruption of ecosystem functions. According to the World Health Organization (WHO) and the European Environment Agency (EEA), long-term exposure to air pollution is associated with reduced life expectancy and increased rates of premature mortality. These concerns are reflected in the Sustainable Development Goals, particularly Target 11.6, which aims to reduce the per capita environmental impact of cities, with a focus on air quality and waste management. The related Indicator 11.6.2 measures the annual mean levels of fine particulate matter (PM2.5 and PM10) in urban areas, weighted by population. In the context of air quality monitoring (SDG 11.6.2 and 13.2.2), advanced earth observation (EO) techniques are employed, integrating multi-temporal Copernicus Sentinel-5P satellite data with deep learning models. This approach monitors pollutants such as NO₂, SO₂, tropospheric ozone, CO, CH₄, CH₂O, and particulate matter (PM2.5, PM10) [3]. These products are supplemented with meteorological data and ground-based measurements. Deep learning models enable precise classification of pollution levels, combining satellite and in situ data for improved accuracy [4]. For studies conducted in Poland, satellite-derived products are validated using national monitoring networks (e.g., the Chief Inspectorate of Environmental Protection), ensuring their reliability [5]. Within this framework, detailed spatial and temporal pollution maps can be developed using GIS and geoinformation platforms. Such maps support urban management, public health planning, and policy-making, particularly in areas with limited ground-based monitoring, contributing to effective and scalable air quality assessment aligned with SDG targets.

Particulate matter with a diameter ≤ 2.5 µm (PM2.5) refers to fine particulate matter with an aerodynamic diameter of 2.5 micrometres or smaller [6]. These particles are small enough to penetrate deep into the respiratory tract and reach the alveoli, posing significant health risks. PM2.5 originates from both natural sources, such as wildfires and dust storms, and anthropogenic (human-made) activities, including vehicle emissions, industrial processes, and residential combustion. Secondary PM2.5 can also form in the atmosphere through chemical reactions involving precursors such as sulphur dioxide (SO₂), nitrogen oxides (NO_x), and volatile organic compounds (VOCs). Due to their small size and chemical composition, PM2.5 particles are a major concern for air quality and public health monitoring.

Particulate matter with a diameter ≤ 10 µm (PM10) refers to a mixture of solid particles and liquid droplets suspended in the air. These particles are small enough to be inhaled and can penetrate the respiratory system, potentially causing adverse health effects. PM10 originates from various sources, including road traffic, industrial processes, and natural sources such as dust or pollen. It is commonly used as an indicator of general air quality, especially in urban environments.

Nitrogen oxides (NO_x) is a collective term for nitrogen monoxide (NO) and nitrogen dioxide (NO₂), two major air pollutants produced primarily during high-temperature combustion processes, such as those in vehicle engines and power plants. NO_x plays a significant role in atmospheric chemical reactions, contributing to the formation of ground-level ozone, smog, and secondary particulate matter. These compounds are harmful to human health, particularly affecting the respiratory system. They also contribute to environmental problems such as acid rain and eutrophication. In air quality assessments, NO_x is a key indicator of combustion-related pollution, especially in urban and industrial areas.

Green areas, addressed in SDG Indicator 11.7.1, are particularly important in densely populated cities, where they contribute to local climate regulation and provide accessible spaces for recreation. In addition to their environmental functions, green spaces support public health and social interaction, which can strengthen neighbourhood ties. As a result, they play a measurable role in enhancing the quality of life in urban environments [7,8,9,10]. Recognizing their significance, there is a growing need to develop comprehensive and up-to-date systems to monitor and assess green areas, enabling informed decision-making for sustainable urban development [11,12,13]. The Sustainable Development Goals (SDGs) focus on creating sustainable cities and communities, emphasizing the importance of green areas, e.g., SDG Target 11.7: “By 2030, provide universal access to safe, inclusive and accessible, green and public spaces in particular for women and children, older persons and persons with disabilities” as an integral component of urban planning and development. Urban green spaces, such as parks and gardens, play a crucial role in enhancing the well-being of urban residents, promoting environmental health, and contributing to overall urban resilience [14,15,16,17]. In the literature, numerous studies are available that utilize remote sensing to assess the availability of green areas within the framework of the Sustainable Development Goals in Poland [18,19]. Within the GAUSS project [7] (Figure 1), a green space assessment system was developed by integrating Sentinel-2 satellite data, vegetation products from the Copernicus Land Monitoring Service, and in situ information. To ensure accurate and reliable validation, the database of topographic objects (BDOT10k) was utilized, enabling detailed characterization of green spaces across Poland. By combining these diverse data sources, the developed service allows end-users to effectively monitor green spaces through the generation of annual statistics accompanied by statistical maps of Poland at the commune level.

The maps presented in this publication were prepared using definitions of green areas derived from Nature Protection criteria. According to the Polish Act [20], green areas, including technical infrastructure and buildings functionally associated with them, are spaces covered with vegetation; those located in the villages of dense buildings or cities; those used for aesthetic, recreational, therapeutic or shielding purposes, in particular parks, lawns, promenades, boulevards, botanical and zoological gardens; children’s playgrounds; historic gardens; cemeteries; and green areas located near roads in and around built-up areas, squares, historic fortifications, buildings, landfill sites, airports, railway stations and industrial buildings. The application of this definition enables the classification of green areas in relation to SDG 11.

2. Related Studies

In order to obtain information on what data and SDG indicators are most often visualized and how, the authors conducted a multilingual query and analysis of publications. An analysis of 120 selected publications showed that only in 91 cases were maps of indicators made in accordance with the cartographic rules. The basic query in Web of Science was as follows: (TS = (map*) or TS = (carto*)) AND (TS = (SDG*)) AND (TS = (Earth*) or TS = (Observation*)). This allowed us to find 221 items. Publications only for SDG 11 and SDG 13 were then selected, leaving 155 results (request link https://www.webofscience.com/wos/woscc/summary/3ed46b7a-bfc0-44ab-a1f1-8da372b8caa4-0176733f08/relevance/1, accessed on 16 June 2025). These items were analyzed. Of them, 53 unavailable in open source, 51 concerning remote sensing methodology, 24 containing no maps, 5 containing maps without a legend, 3 with maps prepared incorrectly (inconsistent with the cartographic methodology), 8 articles on other topics and 2 bibliometric publications were rejected. Ultimately, only nine items were selected that met the assumptions of the analysis. All selected articles were published in English. Based on knowledge of the literature on the subject, 24 articles were added. The next step was a multilingual graphic query. Three graphical searches of the publication were carried out, each time using the following keywords: the full name of SDI, and the words ‘map’, ‘carte’, ‘karte’, and ‘cartogram’, ‘geomorphose’. A total of 87 items was added to the received items that were not included in the WoS but are important from a methodological point of view. In total, 120 items were obtained for the period from 2008 to 2025. Of the 120 publications, 29 were removed, which concerned research results published in other articles and publications containing maps containing methodological errors or lacking a legend. The list of publications containing the 91 items taken into account is available at the following link: https://github.com/AnnMarGeo/article_sdg_methods (accessed on 20 October 2025).

A total of 19 different cartographic methods were applied to visualize the SDG indicators 11.6.2, 11.7.1, and 13.2.2. (Figure 2). Only in five cases were dasymetric solutions [21,22,23,24,25] employed, and in two cases a cartogram [26,27]. Almost all cases were static solutions. Only in one case was a comparison of multiple time states [28] employed to visualize the greenhouse gas emissions in China in the 2000–2011 period. Only one interactive solution was used [29] to present the average air quality map of the greater Cleveland area.

At this point, it is important to define what cartograms are. The cartogram is a map on which a single feature—distance (distance cartograms) or area (area cartograms)—is distorted in proportion to the value of a given phenomenon [30] (see Figure 3). An area cartogram that preserves spatial continuity but modifies the shapes of the basic units in a non-geometric way may be termed an ‘irregular area cartogram’.

It should be emphasized that the use of cartographic methods differs significantly in the case of visualization of the three analyzed indices. For the 11.6.2 indicator visualization, the shaded isoline method is used in more than 43% of cases, e.g., [31,32,33,34,35,36,37,38,39,40,41]. Two cases of its combined use with bars have been identified [42,43] and only one case of combination with a choropleth map [44]. Relatively often (29%), a grid with a different size to the basic field was used as a cartographic presentation method for the visualization of the 11.6.2 indicator [45,46,47,48,49,50,51,52,53]. Sometimes it was limited to the area of communication routes (e.g., [52]). There have been cases of it being combined with choropleth maps (4%, e.g., [50]), as well as with shaded isolines (2%, e.g., [49]). Much less frequent (4%) were cases using quantitative signatures (e.g., [54,55]). One case of the combined use of quantitative signatures with the grid and choropleth method was identified [56]. Relatively rare were cases using the choropleth maps (e.g., [57]), line choropleth maps (e.g., [31]), and cross choropleth maps (e.g., [58]), as well as the combined use of choropleth maps with ordinary signatures (e.g., [59]). Only 6% of the solutions involved the use of chorochromatic maps [60,61,62], 2% of them—grid [63], qualitative signatures [53], isolines (e.g., [64]), and vector maps (e.g., [65]). In most cases, this was used to present the location of green areas in cities.

In the case of the 11.7.1 indicator, the most frequently applied solution is the choro-chromatic method, used in 38% of publications (e.g., [66,67,68,69,70,71,72,73]). A total of 16% are applications of the choropleth method, with units related to city districts (e.g., [74,75,76,77,78,79]. Isochrone (e.g., [80,81,82,83]) and grid (e.g., [84,85,86,87,88]) applications constitute 8% each. Quantitative point signatures are used in 6% (e.g., [89,90]), while buffer (e.g., [91,92]), qualitative point signatures (e.g., [93,94]), and dasymetric choropleth maps (e.g., [21]) are only used in 4% each. The least frequently used methods for 11.7.1 indicators are linear choropleth maps ([95]) proportional symbols (e.g., [96]), shaded isolines (e.g., [97]), and radius maps (e.g., [98]).

The relatively smallest diversity of methods used for visualization occurs in the case of indicator 13.2.2. The most commonly used are choropleth map (e.g., [15,99,100,101]) and grid (e.g., [28,102,103]), at 35% each. In addition, the following methods are used: shaded isolines [104], structural diagrams [105], dasymetric method [106] and cartogram [107], and chorochromatic method [22]. Depending on the aggregation of available data, indicators are visualized in relation to different spatial units.

In the case of 11.6.2 SDG indicator, 50% of maps were elaborated using the point data, which were mostly employed to build the isoline maps (43%). The other 7% of the point data (7%) allowed for the use of quantitative point signatures. A total of 28% of maps employed different grid cells as spatial units. In 2% of cases, the data was related to the NUTS 3. The same percentage was found in the case of city districts and functional areas of cities. In 4.35% of cases, the spatial units were a dense grid of streets. The same percentage was used in attempts to link the index value with land cover elements in cities.

In the case of 11.7.1 SDG indicator, 38% of maps were elaborated using the land use data concerning the cities. In 24% of the cases, the spatial reference for this indicator was the division units. Of these, 10% were city districts and 4% were city functional areas. In 4% of cases, the analysis covered areas within functional areas, while in 6% of cases, the analysis covered areas bounded by streets. A grid with different basic fields was used in 10% of the maps presenting the 11.7.1 indicator.

In the case of 13.2.2 SDG indicator, 45% of maps were elaborated using remote sensing data, concerning the entire Earth. In 25% of cases, the data was related to countries (NUTS 0), in 20% to NUTS 2 and only in 10% to NUTS 3. In 5%, the 13.2.2 SDG indicators were related to the global land use data.

It should be emphasized that, in the case of the analyzed maps of 11.6.2, 11.7.1 13.2.2 SDG indicators, there was no case of referring data from the EO registration to the smallest units of administrative division for the area of a whole country. This solution was used by Panek et al. (2024, [7]).

A critical examination of the relevant literature indicates that cartograms have been employed only infrequently in the representation of sustainable development indicators at the commune and county levels, highlighting a notable gap in spatial visualization practices at finer administrative scales.

3. Materials and Methods

This study examines irregular area cartograms employed for monitoring Sustainable Development Goal (SDG) indicators using Earth Observation (EO) data. The objective is to assess the respective advantages and limitations of this approach in the context of tracking progress toward the SDGs. The prepared maps focus on two key themes aligned with SDG 11 and SDG 13: urban green spaces and air pollution (PM2.5, PM10, NO_x). Both the spatial extent and condition of green areas, as well as the concentrations of selected air pollutants, were derived from Earth Observation data. The applied methodology has been described in the Introduction.

3.1. Study Area

As part of the study, statistical maps were generated to illustrate data for Poland at various administrative levels: at the commune (gmina) level (Figure 4A) and the county (poviat) level (Figure 4B). Poland is a country located in Central Europe, covering an area of approximately 312,696 square kilometres. Its administrative structure is organized into three main tiers: 16 voivodeships (provinces), 380 counties (poviats), and 2477 communes (gminas). This territorial division reflects a decentralized governance model aimed at improving administrative efficiency at the regional and local levels.

3.2. Data Sources

3.2.1. Air Pollution Data

The Copernicus Atmosphere Monitoring Service (CAMS) reanalysis is the latest global dataset of atmospheric composition developed under the EU’s Copernicus programme. It provides three-dimensional, time-consistent fields of key constituents (aerosols, reactive gases, greenhouse gases) at 0.1° (~10 km) resolution with hourly values for 2003–2023, building on earlier MACC and CAMS regional reanalyses. For this study, raster data from CAMS [108] and point data from GIOS (Chief Inspectorate of Environmental Protection) were used, while the borders of LUA were obtained from the Head Office of Geodesy and Cartography (GUGiK) to delimit the analysis area. Air pollution data from both CAMS and GIOS refer to the year 2020.

To obtain yearly averages of air pollutants (PM2.5, PM10, NO_x), datasets from CAMS and GUGiK were utilized. The process involved converting hourly pollution data from NetCDF to TIFF format, resampling to the spatial resolution of 0.01° × 0.01° (~1 km × 1 km), cropping it to Poland’s borders (in a grid with a resolution of 1 × 1 km), and calculating monthly averages for each pixel. These monthly images were then resampled, stacked, and used to compute average pollutant concentrations for local administrative units (LAUs). Finally, yearly averages were calculated from monthly data for each pollutant, combined into a single dataset, and converted into a vector file for analysis and mapping. CAMS data about air pollution was validated against GIOS data. R-squared (R²), mean absolute error (MAE) and mean percentage absolute error (MAPE) were calculated in order to perform validation and evaluate how accurate CAMS data is with respect to air pollution estimated by CAMS over Poland. Validation was performed using monthly and yearly averages from both CAMS and GIOS.

3.2.2. Green Areas Data

The study produced maps focusing on two fundamental attributes of urban green spaces: their distribution and health status. The extent of green areas was delineated using Sentinel-2 data and products from the GAUSS project [6], with a spatial resolution of 10 m. Dynamic World data, based on Sentinel 1 and 2 products, was also used to determine forest coverage. To achieve a more precise delineation of green areas at the commune level, in addition to satellite data, detailed vector boundaries from BDOT10k (Topographic Objects Database at a 1:10,000 scale, for example, parks, green cemeteries, trees within the streets) were used. Sentinel-2 products were also utilized to assess the condition of urban green spaces.

Values of the widely used Normalized Difference Vegetation Index (NDVI) were derived from Sentinel-2 data (10 m resolution) to enable comparison with a novel vegetation index based on High-Resolution Vegetation Phenology and Productivity (HR-VPP) data. These parameters were extracted from Seasonal Trajectories of the Plant Phenology Index (PPI) derived from Sentinel-2 satellite observations. Since growing seasons can traverse years, VPP parameters were provided for a maximum of two growing seasons per year. Data concerning the first increase was used for the analyses. Subsequently, the data on the extent and condition of green areas were aggregated to the commune level (LAUs). Depending on the chosen cartographic presentation method (choropleth maps, or area cartograms), additional statistical data from the Local Data Bank (Central Statistical Office) were incorporated. These supplementary datasets included demographic information, such as population by age groups at the commune level and vehicle counts at the county level.

3.3. Map Production

3.3.1. Colour Legend

An integral component of the map development process was the creation of legends for the choropleth maps illustrating air pollution. Reference ranges for pollutant particle concentrations and their corresponding map colors were adopted from the literature (

Table 1. Air quality classification based on PM2.5, PM10, and nitrogen oxide (NO_x) concentration levels (µg/m³). Based on [6].

Color	PM2.5 Range (µg/m3)	PM10 Range (µg/m3)	Nitrogen Oxide (NOx) Range (µg/m3)	Air Quality Description
Green	0–12	0–20	0–40	Good air quality
Yellow	12–35	20–50	40–90	Moderate air quality
Orange	35–55	50–100	90–180	Unhealthy for sensitive groups
Red	55–150	100–200	180–280	Unhealthy
Purple	>150	>200	>280	Very unhealthy/Hazardous

). These color scales were subsequently adjusted to better reflect the dataset. In the later stages of map preparation, additional subdivisions were introduced within the scales while maintaining the original color scheme; for example, the ‘good air quality’ category was refined by incorporating lighter and darker shades of green. The extended legend is applied to Figures 6–8. Detailed numeric classes and color categories are provided with those figures; the green tonal range is expanded in particular to capture the variability in ‘good’ conditions across Poland.

3.3.2. Choropleth Maps

All choropleth maps were created using QGIS Desktop, version 3.38.0. The underlying base map was the 2020 administrative division of Poland, delineated at the county (poviat) or commune (gmina) level. For all choropleth maps, the ‘natural breaks’ (Jenks) classification method was applied, meaning that class intervals were optimized based on the frequency distribution of the data to minimize within-class variance and maximize between-class variance.

3.3.3. Area Cartograms

This research addresses selected area cartograms, placing special emphasis on automatically generated rectangular cartograms. Various software tools can be used to generate contiguous area cartograms [109]. These may include dedicated software for designing cartograms, plugins or toolboxes for GIS applications, or scripts written in programming languages. A review of selected software tools that support the automated generation of contiguous area cartograms is presented in

Table 2. Software for the generation of contiguous area cartograms. The table presents the type of area cartogram (based on [30]), the name of the software, and a brief summary.

Tools	Software/Language	Cartogram Type	Summary
ScapeToad	Java	Irregular— Gastner-Newman	Desktop application, diffusion algorithm [110]
Cartogram Geoprocessing Tool	ArcGIS Toolbox	Irregular— Gastner-Newman	Implements Gastner-Newman algorithm within ArcGIS environment
RecMap	R	Rectangular or Mosaic	Produces cartograms using rectangular subdivision with attribute scaling [111]
Tilegrams	JavaScript	Hexagonal	Uses equal-sized hexagons or squares; suitable for web presentations (Pitch Interactive)
cartogram 3	Python (PyQGIS) QGIS Plugin	Irregular— Gastner-Newman	Integrates cartogram generation into open-source QGIS environment [112]
cartogram: Create Cartograms with R	R	Irregular—gridded	Is actively maintained and suitable for creating gridded cartograms [113]
go-cart	C++	Irregular—Flow-Based	Create an area cartogram, using Flow-Based-Algorithm [114]

.

The irregular Gastner-Newman cartograms in this publication were produced using the QGIS Desktop plugin cartogram3 version 3.1.5. In the case of communes, two iterations were sufficient to achieve a satisfactory image, while five iterations were used for counties. With fewer iterations, the cartogram adjustment was less accurate. In both cases, the maximum mean error in the anamorphosis process does not exceed 15%.

In this study, the maximum mean error is used as a measure of the accuracy with which statistical values are represented in individual cartogram units. The error is computed as follows: first, a local area error is defined for each unit; next, global measures are derived from these local errors (mean error, maximum error, weighted error, etc.); finally, an error threshold is applied as a stopping criterion for the algorithm (error tolerance). The term “a maximum mean error in the anamorphosis” was adopted from the QGIS software and the plugin that was used in this article to generate the anamorphoses.

There is no widely accepted metric for evaluating irregular cartographic anamorphoses, and quantitative assessment remains limited. As a practical proxy for shape evaluation, the Borderline Accuracy Index (BAI, Equation (1), [115]) known from choropleth map research was used. In the field of cartography, the index is most commonly used for the assessment of choropleth maps, the method of classification, and map complexity. The BAI index expresses the proportion of correctly retained adjacency relations between neighbouring units, compared to the total number of such relations in the reference map. The lower the BAI value, the more distorted the boundaries of the units in the cartogram.

$$\mathrm{B}\mathrm{A}\mathrm{I}=1-{\displaystyle \frac{\mid \mathrm{E}\mathrm{o}\mathrm{r}\mathrm{i}\mathrm{g}\setminus \mathrm{E}\mathrm{c}\mathrm{a}\mathrm{r}\mathrm{t}\mathrm{o} \mid }{\mathrm{E}\mathrm{o}\mathrm{r}\mathrm{i}\mathrm{g}}}$$

(1)

where

Eorig—the set of adjacency relations (shared borders) in the reference map;

Ecarto—the set of adjacency relations in the cartogram;

∣Eorig∖Ecarto∣—the number of adjacency relations lost in the cartogram;

∣Eorig∣—the total number of adjacency relations in the reference map.

The Borderline Accuracy Index (BAI) was calculated using QGIS. The cartograms were first converted from polygon to line features, which enabled the measurement of boundary lengths shared by adjacent units. Information on the BAI value is provided for each of the developed anamorphoses in Section 4, Results.

3.4. Preliminary Usability Survey

To evaluate the usefulness of the elaborated cartograms, a preliminary survey was conducted among potential map users. The survey was made available online via a Google Form and could be accessed at the following link: https://forms.gle/WWhKzdb5cVNPGXnT9 (accessed on 20 October 2025).

The questionnaire addressed several aspects, including the respondents’ level of geographical and cartographic knowledge, their ability to interpret and use choropleth maps and cartograms, and their evaluation of regular cartograms (Figure 5 presents example questions included in the survey). In the pilot survey, nine maps were presented, of which six were cartograms. The maps referred to the territory of Poland, using either the commune or county level as the spatial unit. All questions in the survey were closed-ended, with single- or multiple-choice responses. For one of the questions, respondents could also provide their own answer in a free-text field via the “other” option.

The survey was of an informative and exploratory nature and served as an introduction to a more extensive study planned for the subsequent stages of the project.

4. Results

4.1. Air Pollution

The initial set of maps focuses on air pollution issues related to Sustainable Development Goals (SDGs) 11.6 and 13.2. In the first phase, three choropleth maps were produced (Figure 6), illustrating the spatial distribution of PM2.5 (Figure 6A), PM10 (Figure 6B), and NO_x (Figure 6C) concentrations, respectively, in LAUs in 2020. A uniform color scale was applied throughout, with green hues indicating the lowest levels of air pollution. The analysis of these maps indicates that, in 2020, Poland’s overall air quality was relatively good. However, elevated pollutant concentrations were recorded in major urban areas (e.g., Warsaw, Łódź), within the Upper Silesian Industrial Region (particularly in the case of PM2.5—Figure 6A), and along key transportation corridors (especially NO_x, notably the belt extending from Warsaw and Łódź toward the Upper Silesian region—Figure 6C). According to the spatial patterns presented on the maps, the areas with the lowest air pollution concentrations are located in northern Poland, especially along the Baltic Sea and in proximity to the Kaliningrad and Lithuanian borders, as well as in the mountainous regions of the south, such as the Bieszczady Mountains and the Tatras.

The second step involved producing area cartograms to depict PM2.5 (Figure 7A) and PM10 (Figure 7B) concentrations within Polish communes in 2020. The area of each commune on the cartograms is scaled proportionally to its population. The mean cartogram error was 15.6%, while the BAI index amounted to 0.69. This approach was chosen because population size directly relates to emissions, as residential sources constitute the main contributors to both PM2.5 and PM10 pollution.

It can be observed on the cartograms that larger commune areas generally correspond to lighter shades of green or yellow. The use of cartograms distinctly highlights communes with large populations but small geographic areas and their associated pollution levels. On cartograms, major cities become distinctly visible, including Szczecin, the Tricity metropolitan area (Gdańsk), Poznań, Łódź, Warsaw, Wrocław, the Upper Silesian Industrial Region, and Kraków, roughly arranged from northwest to southeast. In contrast, these areas were less clearly represented in the previous standard choropleth maps (Figure 6), where cities like Szczecin, the Tricity, and parts of the Upper Silesian conurbation appeared less prominent.

In the NO_x concentration cartogram, the territorial division of Poland into counties (poviats) was used, with the area of each unit scaled according to the number of vehicles registered in that county in 2020 (Figure 8). The mean cartogram error was 8.6% (reflecting the statistical representation of the data), and the BAI index amounted to 0.74. Applying an area cartogram in this context produced a notably informative outcome: counties with the largest visual representation—reflecting the highest number of registered vehicles—are consistently shaded in the lightest green, corresponding to the highest NO_x concentrations. This spatial pattern is not as easily discernible in the standard choropleth map (Figure 8A). By using the area cartogram (Figure 8B), NO_x concentrations in the Tricity area became clearly discernible, something that was challenging on the choropleth map owing to the small size of the constituent urban counties.

Nevertheless, it is important to acknowledge that the application of area cartograms can result in considerable spatial distortions. This is especially evident in counties with large territories but relatively few registered vehicles, which can complicate the interpretation of data in regions such as Eastern Poland. In the case of larger shape distortions, it may be useful to present cartograms alongside an administrative map. The most effective solution would be to provide cartograms in the form of a web application/geoportal with an interactive panel displaying, among other things, information on the name of the administrative unit.

4.2. Green Areas

The choropleth map (Figure 9) and area cartograms (Figure 9 and Figure 10) were generated for green areas to illustrate both their coverage and overall condition. Figure 9 presents the extent of green areas using a choropleth map (Figure 9A), as well as both the extent and condition of green areas using a cartogram (Figure 9B). The mean cartogram error was 17.9% (reflecting the statistical representation of the data), and the BAI index amounted to 0.63. In the choropleth map, increasingly intense shades of green indicate a higher proportion of urban green space relative to the total commune area. In the cartogram, the intensity of the green shade corresponds to the condition of green spaces; the stronger the hue, the better the ecological state.

It is evident that in certain communes, including Warsaw and the Tricity, green areas represent a substantial proportion of the total commune territory; however, their ecological condition ranges from moderate to poor. Such nuanced information cannot be as effectively discerned from the choropleth map alone.

In relation to Target 11.7—‘Provide access to safe and inclusive green and public spaces’—a critical consideration is the assessment of accessibility for specific demographic groups, particularly children and the elderly. Figure 10 presents a set of cartograms illustrating both the extent (Figure 10A) and ecological condition of green areas (Figure 10B) in correlation with the number of children (aged 0–14) residing in each commune. The mean cartogram error was 6.9%, while the BAI index amounted to 0.61. Findings indicate that in communes with relatively large child populations, the ecological condition of green spaces is not necessarily high. The use of area cartograms enhances the interpretive capacity of spatial analyses by simultaneously representing accessibility and environmental quality. This approach underscores the critical role of green space quality in supporting the health, development, and well-being of younger demographic groups.

4.3. Preliminary Usability Survey Result

A pre-survey on the potential use of irregular area cartograms was completed by 19 respondents (12 male, 7 female). The majority indicated intermediate or high domain knowledge in geography and cartography (16 of 19). An initial skill check embedded at the start of the instrument confirmed these competencies. Respondents comprised researchers (not academic teachers—42%), public administration employees (26%), academic teachers (11%), non-governmental organization employees (5%), and other respondents (16%).

Analysis of the pre-survey items suggests that irregular cartograms are well-suited for presenting data to detect spatial patterns and communicating ancillary information (Figure 11). However, the study also shows that respondents more often rely on the traditional choropleth when reading data for a single spatial unit.

One objective of the survey was to select problem areas where irregular area cartograms could be meaningfully employed. The preliminary shortlist comprised air pollution and green-space indicators at the municipal level. Post-survey analysis indicated that further relevant domains may include the following:

• Proportion of population that has convenient access to public transport (12 answers);
• Proportion of municipal waste generated according to the treatment operation to total municipal generated (11 answers);
• Total greenhouse gas emissions per year (9 answers).

These themes will be pursued in subsequent research efforts.

In the pre-survey, respondents (n = 19) provided a subjective appraisal of irregular area cartograms across eight aspects (Figure 12), using a four-point scale (1–4). Overall, irregular area cartograms were rated as interesting, informative, well-organized, and effectively conveying the map’s underlying idea. The lowest scores were assigned to map readability, indicating persistent interpretability challenges. Notably, the method was also judged not entirely trustworthy, despite communicating less-generalized information than traditional choropleth maps. This divergence suggests that some participants may not have fully internalized the methodological premises of area cartograms, which in turn may have affected perceived credibility and ease of use.

Based on a preliminary survey among potential users of area cartograms, this method can be considered effective for conveying new information on the map and for rapidly detecting spatial patterns. However, it is still regarded as difficult to read and less intuitive than a traditional choropleth map.

5. Discussion and Conclusions

The analysis carried out has shown that despite the growing importance of spatially disaggregated data in sustainability assessments, cartographic research has seldom addressed the visualization of SDG indicators at detailed territorial scales. The cartographic methods most frequently applied in the context of the Sustainable Development Goals (SDGs) are choropleth and grid maps, which is largely due to the availability of these methods in the software used by the authors. These visualizations, however, are typically limited to national or broad regional levels of analysis. Analyzed maps were used only to visualize simple spatial information. Only in a few cases, e.g., [31,81], was a complex map (presenting interconnected phenomena) employed. The remaining authors preferred to present several maps on different topics, e.g., [73,91,98]. No analytical complex maps were used that could show related phenomena and their causes. Creating such maps would require the use of more advanced cartographic software and (at least) consultation with cartographers. Among the authors of the examined sample of articles were only five cartographers, which unfortunately affects the methodological level of the maps and the possibilities of perceiving the phenomena presented on them.

The analysis of maps in the articles, conducted from the point of view of a cartographer, showed that only 56% of them were made relatively correctly. In the case of indicator 11.7.1, authors often employ extracts of green areas from topographic or thematic databases, e.g., [40,77,85,90]. Less frequently, they use maps based on the NDVI satellite data index, e.g., [66]. Several cases of maps without legends were identified. In one case [45], a cross legend was used without any description of the values, rendering the map useless. In many cases (of all indicators), maps lack legends with numerical values. They are replaced by information such as ‘high’, ‘medium’, and ‘low’, e.g., [31,33,42,51,77,80,89,97,101]. It is common to incorrectly use continuous legends to present data in intervals or assign a single value to the entire interval. In a few cases, continuous choropleth map legends were used to describe a discrete choropleth map. In seven cases, the choropleth map legend was used to describe isolines [31,34,36,38,39,40,41], while in six cases, an isoline legend was applied to a choropleth map, e.g., [44,52,58,65,99,100]. In one case [45], a choropleth map was used to present absolute values (which is considered a serious error). In 80% of the maps examined, the choropleth map legend scale was upside down, e.g., [25,31,37,57,77,78,83,87,90,99]. In the study group, only 12% of the choropleth maps had correct legends. In nearly 40% of the maps, the principle of using ‘warm’ colors to represent high or positive values and ‘cold’ colors for low or negative values was not followed, e.g., [31,46]. This, in turn, leads to incorrect information transfer and user perception. The significant scale of neglecting the rules of cartographic methodology in this period when AI is based on available publications and is not yet perfect in the case of spatial data may cause the risk of duplicating and multiplying errors. The problem of the methodological correctness of the visualization of SDG indicators has already been signaled in many reports, e.g., [116,117,118,119]. A scientific discussion on this topic could constitute material for a separate publication. The existing needs have led to several attempts to create tools to facilitate the visualization of SDG indicators, e.g., the ‘Sustainable Development Goals Interlinkages and Indicators’ project of IGES [120] and the more developed (and methodologically correct) SDG Viz Web-based System [121]. It should be stressed, however, that these solutions did not use cartograms. These map-creating methods are widely (but wrongly) considered difficult. It is true that not every software can create them. However, it is important to emphasize that these methods allow for the simultaneous visualization of the state of phenomena and their determinants. Their use in visualization also allows for deducing the causes of the scale of phenomena presented by the SDG indicators.

The synthesis of existing academic research (Section 2: Related Studies) demonstrates that the use of irregular area cartograms in representing Sustainable Development Goals (SDGs) remains exceedingly limited, especially in studies that employ Earth Observation data. Moreover, where such techniques have been applied, they have largely been restricted to aggregated territorial units, with little attention given to subnational levels such as counties or communes. A notable exception is the publication Mapping for a Sustainable World [122], where Section 3.8 briefly explores the potential of area cartograms for visualizing SDG-related data at the national scale.

Analysis of the literature, combined with the cartographic outputs developed in this study, highlights the practical value of using irregular area cartograms to visualize Sustainable Development Goal (SDG) indicators. This method is particularly useful in representing urban areas, which are often visually minimized in traditional cartographic techniques such as choropleth maps due to their small geographic size. This issue is especially relevant in the Polish administrative context, where rural communes typically cover much larger areas than urban units, including cities with county status.

Despite these advantages, certain limitations of irregular area cartograms must be acknowledged. Irregular or highly distorted cartograms may impede the legibility and spatial recognition of individual administrative units, particularly for end-users without prior familiarity with the geographic configuration of the study area. Such distortions can pose challenges to interpretability and reduce the communicative effectiveness of the map.

The results of this study indicate several strategic directions for advancing the application of irregular area cartograms in the spatial visualization of SDG-related phenomena, particularly in contexts involving continuous Earth Observation (EO) data:

• Applying irregular area cartograms to represent spatial units with highly variable population sizes at lower administrative levels (e.g., communes, counties). This technique enhances the visibility and interpretability of densely populated urban areas, which are often spatially limited but demographically significant;
• Integrating Earth Observation data into the construction of irregular area cartograms, which enriches the thematic content of the maps and enables more frequent and dynamic monitoring of urban environments compared to conventionally collected statistical datasets. EO-based inputs offer higher temporal resolution and spatial consistency, supporting timely assessments of sustainability indicators;
• Combining irregular area cartograms with other cartographic techniques, such as choropleth maps, proportional symbols, or qualitative and quantitative point signatures. Such hybrid visualizations provide a more comprehensive representation of SDG-related issues by simultaneously conveying multiple dimensions of the data;
• According to an initial survey among potential users, irregular area cartograms help communicate additional information on maps and facilitate swift recognition of spatial patterns. Even so, they continue to be considered difficult to interpret and less intuitive than traditional choropleth maps.

The testing of the generated irregular area cartograms showed that users are able to assess the scale of the presented phenomena on their basis, as well as draw conclusions about their conditions and causes. The conducted experiment demonstrated the high usefulness of the tested method for visualizing and analyzing SDG indicators. The proposed solution is operational, and in the future, it may be used for more extensive SDG research (also based on EO data from future Copernicus satellites), concerning other countries or regions of the world. It can also contribute to the international scientific discussion on the possibilities and limitations of using cartograms to visualize and analyze SDG indicators.