Urban air pollution remains one of the most critical environmental health challenges in large metropolitan areas, particularly in Latin America. Traditional monitoring systems, while reliable, are often limited in spatial and temporal resolution, leaving gaps in the detection of local pollution events. This study presents a data-driven approach to analyze citizen complaints related to air pollution in Mexico City by leveraging unstructured data from Twitter. A semi-automated pipeline was developed for the extraction, cleaning, transformation, and analysis of over 500,000 tweets collected between 2019 and 2022. After preprocessing, 69,739 tweets were subjected to topic modeling using Latent Dirichlet Allocation (LDA), enabling the identification of twelve semantic categories associated with pollution events—such as traffic congestion, fire outbreaks, fireworks usage, and emergency response. Spatiotemporal analysis revealed that the boroughs of Miguel Hidalgo, Cuauhtémoc, and Iztapalapa recorded the highest number of reports, particularly during festive periods and dry seasons in Mexico City. The results highlight the strong potential of social networks as complementary sources of environmental information, capable of capturing citizen perception, behavior, and real-time local incidents often excluded from institutional air quality datasets. A Power BI Geo-dashboard was developed to facilitate interactive visualization and exploration of pollution-related complaints, supporting evidence-based urban decision-making. This study contributes a replicable framework for social media–based environmental monitoring and advocates for the integration of citizen-generated data into urban air quality management strategies.
Volunteered Geographic Information (VGI) images are a vital source of visual information and image features in the GIS field. The advancement of artificial intelligence, particularly large language models and generative AI, now enables the generation of seemingly lifelike images from textual prompts (Artificial Intelligence Generated Content - AIGC). This raises a pertinent question: can AIGC image features serve as a viable alternative to VGI features in downstream GIS tasks, especially where VGI is scarce or difficult to obtain? This paper conducts an exploratory study comparing VGI and AIGC image features as inputs for a geographic recommendation model, specifically investigating the impact of image elements, colors, and spatial structures. The results indicate that current AIGC images, generated from general textual descriptions, cannot fully substitute for VGI images. This is primarily due to AIGC's challenges in accurately replicating the specific elements, colors, and spatial relationships inherent in real-world VGI. However, the study suggests AIGC images hold significant potential. When provided with more specific information about elements and particularly their colors, AIGC's performance approached, and in some color-focused tests, even slightly surpassed that of VGI images. This implies AIGC's main deficiency is its current understanding of real-world object characteristics and their visual representation, notably the basic knowledge of elements and their associated colors. We propose these shortcomings could be addressed by integrating geographic knowledge bases in future AIGC development. These findings aim to guide AIGC's application in GIS by identifying current limitations and areas for focused improvement.
The exponential growth of tourist-generated content necessitates efficient quality assessment frameworks to address inherent challenges of information reliability and analytical scalability. This study pioneers a Large Language Model (LLM)-driven approach, integrating supervised fine-tuning with low-rank adaptation and structured prompt engineering to enable multi-dimensional quality evaluation. Covering 485,930 reviews from three major platforms-MaFengWo, TripAdvisor, and Ctrip-the framework achieves superior performance (RMSE = 0.56, NDCG@K = 0.88) in generating accurate quality scores and detailed analytical rationales. Spatial-temporal-semantic analyses reveal platform-specific quality patterns: MFW exhibits stable temporal cointegration and prominent spatial centrality, TripAdvisor demonstrates simplified core-periphery structures, while Ctrip presents dynamic multicentricity. Heterogeneous network analysis further identifies the behavioral regularities of high-reliability users through a randomwalk algorithm. The study advances tourism informatics by resolving scalability limitations of manual coding while providing actionable insights for platform governance, including targeted moderation and incentive mechanisms. This paradigm highlights LLMs' transformative potential in operationalizing tourist-generated content quality assessment at scale, bridging theoretical rigor with practical applicability within digital tourism ecosystems.
This research proposes an integrated recommendation framework to support automobile purchase decisions by aligning consumer preferences with spatial and technical factors. In response to the challenges faced by car buyers, particularly those with limited automotive expertise and overwhelmed by excessive information, the model leverages sentiment analysis from social media and combines it with spatial variables, including dealership density, vehicle availability, circulation restrictions, environmental suitability, and pricing. These factors are then matched with technical specifications to generate personalized and context-aware vehicle suggestions. Using Mexico City as a case study, the empirical analysis reveals that regional differences in infrastructure, regulatory policies, economic conditions, and environmental constraints strongly shape vehicle preferences. The findings show that consumers in areas with stricter vehicle regulations and better service infrastructure tend to prefer vehicles with higher initial costs but lower long-term maintenance demands. While the system is built around the socio-economic and infrastructural peculiarities of Mexico City, it has been designed with modular and configurable components, allowing it to be adapted to international contexts with similar data availability. Nevertheless, the dependence on locally specific variables should be explicitly acknowledged when considering its broader applicability. This raises the question of scalability and transferability, inviting further research into how the framework can be generalized across diverse urban and regulatory environments.
Water scarcity in Mexico City has become an increasingly urgent issue, exacerbated by inefficient and unequal consumption patterns across its urban fabric. This study advances Geographic Information Systems (GIS) research by developing and applying an integrative spatial analysis framework specifically tailored to the complexities of urban water management. Beyond its application to Mexico City, the research demonstrates how GIS can be used to fuse heterogeneous datasets, including those from SACMEX (Mexico City's Water System), INEGI (National Institute of Statistics and Geography), and DENUE (National Directory of Economic Units), into a unified analytical environment. Through a combination of exploratory data analysis (EDA), spatial data mining, and clustering techniques, the study identifies critical disparities in water consumption at multiple spatial scales, from boroughs to neighborhoods. A key contribution is the implementation of a layered system architecture for managing historic spatiotemporal data, enabling dynamic visualization of consumption patterns. The findings reveal that socio-economic and demographic variables play a decisive role in shaping spatial water demand, with marginalized communities facing disproportionate challenges. While previous spatiotemporal analyses of water consumption in Mexico City have primarily focused on aggregated borough-level data or isolated socio-demographic correlations, they have often lacked multiscale integration, high-resolution neighborhood-level analysis, or interactive visualization tools to support policy development. This research addresses these limitations by providing a fine-grained, multilayered analytical approach that enhances the scientific understanding of urban water use. Beyond offering immediate policy-relevant insights for Mexico City, the methodological framework proposed here contributes to GIS research by providing a scalable, transferable approach for analyzing urban resource consumption patterns. Future work will focus on incorporating real-time data streams, expanding sector-specific analysis, and integrating additional variables and domains required for a comprehensive understanding of water dynamics. This study represents a first step toward building an adaptive, equitable, and efficient urban water management strategy.
Small and medium-sized enterprises (SMEs) are the backbone of Mexico’s economy; however, they frequently face high failure rates due to suboptimal location choices and limited access to market information. In this study, we introduce a geospatial framework for predicting the longevity of entrepreneurial SMEs in Mexico City, using veterinary clinics as a representative case. Our approach integrates multiple datasets—including socioeconomic, geographical, and criminal activity records—to construct an extensive feature set capturing neighborhood demographics, commercial surroundings, and urban infrastructure. We then apply supervised learning methods, including multilayer perceptrons, random forests, and gradient boosting classifiers, to classify the expected lifespan of each business establishment into three survival groups (1–2, 3–4, and 5–6 years). Experiments with cross-validation reveal that an optimized XGBoost model achieves a balanced accuracy of 88.68
Recent advancements in airborne platforms equipped with ultra-high-resolution imaging sensors have significantly improved our capability to acquire detailed urban imagery. These systems offer exceptional clarity and precision in capturing urban data, paving the way for innovative approaches to urban analysis. However, manually extracting information from this data can be a slow and labour-intensive process. Thus, employing deep learning algorithms for data extraction appears to be a promising solution. However, deep learning methods require significant amounts of training data, processing power, and memory to achieve satisfactory results. To tackle these challenges, we developed a new approach called the Color Range Mask (CRM) layer, which facilitates an in-depth understanding and assessment of input images. This method enhances the effectiveness of CNN-based deep learning algorithms for image processing while maintaining computational and memory efficiency. We evaluated our approach using both U-Net and pre-trained ResNetUNet deep learning architectures. The CRM layer proved to be effective, slightly improving segmentation accuracy, particularly for complex classes such as Crosswalks and Water bodies, which have a limited amount of training data. Comparative assessments revealed that pre-trained ResNetUNet models incorporating the CRM layer outperformed baseline U-Net models across various metrics, including Intersection over Union (IoU) and Accuracy.
Mexico City has more than nine million inhabitants living in marginalized regions characterized by high crime rates and significant air pollution. According to data provided by the Government of Mexico City, the areas with the highest reported crime rates since 2014 are located in the central, northern, and eastern parts of the city; these same areas also record historically high levels of fine particulate matter (PM10 and PM2.5). Current studies are focusing on the psychological dimension and the possible link between air pollution and criminal behavior, suggesting that ongoing exposure to elevated levels of PM10, CO, and ozone may lead to cognitive deterioration, increased anxiety, depression, and aggressive conduct. In this context, how can we integrate, analyze, and explore crime and air quality data to identify correlations between these phenomena? This paper proposes an advanced urban data analytics framework that outlines phases for integrating datasets and detecting correlations between rising air pollution and the crimes recorded at the borough level in Mexico City. Using data mining techniques, we integrated and analyzed trends in PM10, PM2.5, CO, ozone, and both violent and non-violent crimes from 2016 to 2020, drawing on open data sources from the city. We included records of crimes such as robberies with violence, theft in public transportation and urban spaces, domestic violence, and homicides. These data were examined in greater depth using Pearson correlation and linear regression. The results indeed suggest a positive and statistically significant relationship between boroughs with higher crime rates and elevated air pollution.
The development of smart and healthy cities requires the creation of urban frameworks that comprehensively assess safety levels in transportation systems. Mexico City’s subway system (Metro) is the eighth most traveled in the world, while the city’s Bus Rapid Transit system (Metrobús) transports approximately 1.8 million passengers daily. However, the Metro has been rated as one of the least secure electric transport systems, underscoring the need for effective strategies to mitigate safety risks. This study proposes a framework based on artificial intelligence and open data to assess security at Metro stations and Metrobús stops. Crime records from 2019 to 2022, georeferenced transport network layers, and passenger flow data were analyzed to identify crime patterns within public transportation environments. Findings reveal that women are the primary victims of harassment and sexual abuse, whereas men are more frequently targeted for street robberies and passenger thefts inside Metro trains. Additionally, crime rates peak at noon, with the most vulnerable age group being individuals aged 21 to 30. Notably, half of the recorded Metro-related crimes occur within the influence zones of Lines 1, 2, and 3. Based on these insights, a predictive model using SARIMA and Support Vector Machines (SVM) was developed to estimate weekly crime risk at transport stations and their surrounding areas. The SVM model achieved 70
Understanding the geospatial and temporal distribution of diabetes mellitus in Mexico can be an essential tool in supporting vulnerable populations and addressing health inequalities. This article presents a spatio-temporal investigation of patients aged 18 years and older with diabetes mellitus in Mexico, associated with geographical area and a temporal range from 2005 to 2022. This approach includes calculating diabetes-related hospitalizations and deaths and its association with the margination index segmented into eight geographical areas of Mexico. Furthermore, this research stratifies based upon age group and type of medical institute of the health services in Mexico. The main contribution of this research is to explore the relationship between diabetes-related hospitalizations, deaths, geographical area, age, sex, and margination index of populations to support preventive action. The results highlight that adults between the ages of 45 and 64 years old who live in areas with a high margination index have a greater likelihood of suffering complications related to diabetes. The age-adjusted rate of DRAH shows that the Peninsula has the highest values among geographical areas. Research will now continue to explore mapping interventions to specific states and external datasets, to further extrapolate the results of the analysis.
Natural disasters pose a constant threat to urban populations, causing severe risks to human safety, infrastructure, and the environment. The United States experiences approximately 50 major disasters annually, including hurricanes, wildfires, and floods, often leading to significant challenges in emergency management. Traditional response systems face limitations such as information overload and coordination issues, especially in densely populated areas. To address these challenges, this article explores the integration of Artificial Intelligence (AI) and GIScience to enhance real-time decision-making during crises. This work introduces a conversational virtual assistant based on ChatGPT-4o, designed to process real-time geospatial data and assist citizens during emergencies. The system utilizes Large Language Models (LLMs) and geospatial resources to access, synthesize, and analyze official sources, generate customized evacuation routes, and provide actionable insights. A flood emergency case study is employed to illustrate the system's potential to transform urban disaster response.
Geographical Information Systems (GIS) are essential tools used for storing and performing analyses on spatio-temporal data. GIS reporting plays a crucial role in transforming the result of these analyses into actionable insights, enabling informed decision-making, by uncovering patterns and relationships and presenting them in a human-readable format suitable for the intended target audience. Nonetheless, the traditional process of creating reports involves manual analysis and interpretation of spatio-temporal data, a time-intensive task prone to human error. This paper aims to investigate the potential of Large Language Models (LLMs), particularly their natural language processing and generation abilities, to streamline the report generation process. To this end, three case studies are conducted, using the GPT-3.5 LLM to analyze real-world GIS data, extract key spatio-temporal insights and generate actionable, human-readable reports. The generated reports are then analyzed, to assess the model’s capacity for understanding complex spatio-temporal relationships and patterns and generating coherent reports. Results show that general-purpose LLMs can be remarkably effective in detecting spatio-temporal patterns and anomalies and in generating concise, effective human-readable reports. Despite this great potential, we also identify several key challenges of LLMs for GIS report generation, including a significant variability among different re-executions, a tendency to report incorrect data in some scenarios, and difficulty in understanding more complex spatial data such as polygons.
Toponym resolution can be defined as the process of mapping each toponyms in a document or corpus unambiguously to an associated spatial footprint corresponding to the location intended by the writer of a document. Various toponym resolution methods have been proposed that use the latitude and longitude of the centroids as the spatial representation, but arguably large cities regions or countries are poorly represented by a single point. Two decades ago, Leidner presented TAME, the first geo-annotation tool described in the literature, but it (a) lacked extensibility (such as support for multiple users and multiple gazetteers), (b) it is not publicly available, and (c) it did not support polygon footprints. As a consequence, to date, in all annotated text collections, place names are associated exclusively using gazetteers with centroid geographic footprint representations. In this paper, we present TAME II, a more flexible system for creating labeled corpora for the training of and evaluation of toponym resolvers, which supports multiple kinds of geographic footprint types (polygons, centroids and bounding rectangles. It has been implemented as a modern Web-based application that is available for the public on the Web. To the best of our knowledge, TAME II is the first text annotation tool that supports gazetteers with polygon footprints, and the first Web-based tool generally available.
Urban landscapes are rapidly evolving, integrating diverse Points of Interest (POIs) to accommodate city dwellers’ needs, highlighting the necessity for efficient analytical frameworks. This study presents a preprocessing framework using KNIME, known for its user-friendly interface and robust data management, for efficient POI preprocessing. We integrated the Advan mobility dataset with the Census dataset, allowing us to consider both the POI features of the location and the characteristics of the people using these POIs. We also introduce two novel POI features that link visitors to POIs: median dwell time at POIs and visitor travel distance from home to POI, to deepen our understanding of POI dynamics. We used the framework to conduct spatiotemporal analyses across 31 POI categories, identifying significant temporal variations linked to daily human behaviour within these categories. For our spatial analysis case study, due to healthcare disparities across various geographic divisions in the US, we selected the outpatient care services category of POIs for deeper analysis. The study underscores a significant correlation between our two novel features and variances in both geographic (land area) and demographic (population density) aspects across nine US divisions. This research makes a substantial contribution to urban studies, providing a solid framework for POI analysis and introducing other influential features along with visitors’ check-in data for examining POIs in cities.
User-generated geospatial data have opened up new opportunities for a wide range of applications. In particular, civic monitoring initiatives based on volunteered geographic information (VGI) have allowed ordinary people to mobilise Earth Observation technologies for environmental monitoring to detect non-compliance with environmental laws and regulations. Through review of illustrative cases, this article will show how such data have played a key role in triggering investigations by public prosecutors and are grounding claims in front of the judiciary by affected citizens. The article will also discuss cases where civil society used geospatial data to demonstrate non-compliance with environmental rules, ban extractive operations and demand reparation for environmental damage.
Urban Functional regions represent dynamic formations molded by spatial interactions, illustrating the intricate connections between different geographic areas. Identifying functional regions is essential for supporting urban planning efforts and promoting sustainable development. This study presents a comprehensive framework to characterize critical road network locations, integrating structural, functional, and geographical dimensions of the built environment. Addressing limitations in one-dimensional networks and multivariate issues, the study utilizes spatial density-based hotspot detection approaches and develops a method to identify critical locations as multivariate hotspots. Emphasizing critical locations as foundational units for functional regions, it then addresses spatial interaction data limitations through flexible modeling approaches, enhancing local modeling with Artificial Neural Networks (ANNs) within Geographically Weighted Regression (GWR) models. Lastly, dynamic spatial interactions are considered using overlapping community detection methods, offering a structured framework for identifying overlapping functional regions in urban landscapes. This study enhances the identification and analysis of urban functional regions, providing deeper insights into spatial complexities and facilitating more effective decision-making for sustainability.
Most common GIS applications are composed of a spatially enabled database and a desktop client directly connected to it. In this configuration the business logic is often found either in the client application with custom plugins or in the database where many triggers are defined. When it comes to web GIS applications, a cartographic server is added in between the database and a web client. In this paper, we present an alternative setup where the data-model and the business logic are both defined in Python programming language using the Django web framework ORM. Both desktop and web clients exchange data through OGC services. In addition, the business logic is implemented in a middleware in a Django application. While mitigating the issues we face in standard PostGIS-based solutions, the Django ecosystem also comes with powerful tools offering interesting perspectives for such applications.
Despite its soil protection regulation, ongoing land sealing and soil loss have continued over the past three decades and Switzerland lacks a comprehensive soil map. Soil quality is affected by land use which is challenging to map because it may change on the short-time, and also soil functions are controlled by soil properties that are not necessarily taken into account in soil mapping, such as organic matter content. Moreover, stakeholders seldom integrate soil information in their projects, since the data is scarce and the soil awareness is poor. To address these challenges, we developed a user-oriented framework assessing the soil functions based on expert predictions from land use, which are corrected with available soil data. The framework was applied to estimate soil quality in the Morges region. It takes into account all available digital information such as land-use layers, aerial imaging, geological maps etc. to split the area into polygons of predicted soil quality associated with an estimated accuracy, and integrates in a second step all available information on the soils, such as soil analysis, soil depth, polluted sites etc. Two complementary tools based on open-source software allow refining the estimated scores of the polygons and simulating the impact of development projects on soil quality. A semi-automated process was implemented to model input parameters and generate maps. Workshops involving experts, decision-makers, and soil specialists from the region allowed to improve and validate the soil mapping and the interface. The framework, named IQSM, allows to generate and update maps inexpensively, this can be replicated in any Swiss location. The open-source methods and the user-oriented interface raised a strong interest among stakeholders.
Shipping companies play a crucial role in coordinating container transportation, as their decisions about shipping routes and ports have a direct impact on the development of the container network. Hence, understanding the similarities and distinctions across ship companies is crucial for gaining insight into the container shipping market. This study uses a multiplex network method to provide insight into identifying various transportation patterns and significant ports. The model is enhanced with a set of structural indices, together with a novel overlap measure that assesses the distinct function of a particular port across several layers of ship companies. The entire methodology is applied and tested using AIS data from the top 10 shipping companies, which facilitates the creation of a multiplex network and the calculation of structural metrics. The experiments conducted on the global maritime transportation network aim to identify crucial ports and emphasize notable variations at both the regional and trade flow levels. This study enhances the existing knowledge in the field of transport geography and uncovers the route organization strategies employed by shipping companies. These findings can offer valuable insights for port managers when making decisions.
This paper addresses the intricate phenomenon of missing persons, a major problem in Mexico, extending beyond the simple occurrences of disappearances. By developing a historical and spatial data analysis, this research comprehensively examines missing person data from open, official, and social media sources. We apply a data mining framework based on digital media and openly accessible government databases to characterize and visually represent crimes such as enforced disappearances along the temporal and spatial dimensions. The data analysis methodology takes a comprehensive approach, segmenting data by age, sex, nationality, geographic location, and period. This segmentation unveils patterns in space and time, thus contributing to a better understanding of the factors influencing missing person phenomena and valuable insights into the dynamics of missing person incidents that have impacted many states and regions in Mexico over the past decade.
