Back to Programme

Session 32

Geoinformatics and Spatial Intelligence: Methods, Applications and Challenges

Organisers: Dimitris Kavroudakis (University of the Aegean)

  • Friday, 27 November 2026 · 09:30–11:00, 11:30–13:00, 16:30–18:00, 18:30–20:00

    Room 5, Harokopio University of Athens (HUA)

About this session
Geoinformatics has become a core pillar of contemporary geographical research, enabling the collection, integration, analysis, and visualization of spatial data across multiple scales. Advances in geographic information systems (GIS), remote sensing, spatial data infrastructures, open geospatial data, and computational methods are transforming how geographers model, analyze, and interpret complex spatial and environmental processes. This session aims to bring together contributions that explore theoretical, methodological, and applied dimensions of Geoinformatics, with particular emphasis on technological developments and analytical workflows. The session focuses on the role of geospatial technologies in addressing key geographical challenges related to climate dynamics, urban systems, environmental risk, and territorial organization. Special attention is given to the integration of GIS and remote sensing with data science techniques, spatial modeling, machine learning, automation, and decision-support systems. The session welcomes empirical case studies, methodological innovations, and educational perspectives that demonstrate how geoinformatics supports advanced spatial analysis and reproducible research. Contributions may address topics such as spatial data quality and uncertainty, multiscale and spatio-temporal analysis, geovisualization, smart and sensor-enabled environments, environmental monitoring, spatial planning tools, and the use of open-source software and open-data ecosystems. By fostering dialogue between physical and human geography, applied sciences, and spatial planning, this session underlines the relevance of Geoinformatics as a bridge between geographical theory and technological implementation. In line with the conference theme "Geography Matters", the session highlights how spatial intelligence and geospatial technologies contribute to more robust, transparent, and methodologically sound geographical research and decision-making.

Presentations are currently listed in alphabetical order (by first author). The final order of presentations has not yet been confirmed and will be announced soon.

Presentations18

  1. Assessment of road network connectivity using GIS-based spatial analysis to enhance organized preventive evacuation planning.

    Panagiotis Agourogiannis

    • University of the Aegean
    Abstract
    The climate crisis and the escalating frequency of natural disasters necessitate a comprehensive reassessment of civil protection strategies, especially in regions facing demographic pressure driven by tourism in both urban and insular environments. This influx of visitors, coupled with spatial heterogeneity, exacerbates the challenges of timely and efficient evacuation during emergencies such as seismic events, wildfires, or floods. Consequently, the formulation of targeted preventive measures and adaptive evacuation scenarios has become more critical than ever. This study leverages GIS capabilities and spatial network analysis to evaluate road topology as a pivotal factor in evacuation efficacy. The research aims to establish an integrated geospatial framework for assessing accessibility to emergency shelters, grounded in fundamental metrics such as intersection density, clustering coefficients, link-node ratios, and the distribution of Points of Interest (POIs). Utilizing spatial modeling within a GIS environment, the study investigates how intersection density modulates evacuation dynamics by offering redundant routing options. Concurrently, clustering analysis identifies zones of high local connectivity that bolster resilience at the neighborhood scale. Furthermore, the interplay between links and nodes is employed to map network permeability, pinpointing critical bottlenecks and culdesacs that impede the flow of civilians and emergency response units. The application of GIS facilitates the precise spatial pinpointing of network vulnerabilities and their spatial correlation with shelter locations. The findings indicate that geospatial connectivity analysis enables dynamic population rerouting and the prioritization of structural interventions. The study culminates in specific policy recommendations for refining civil protection protocols, proposing a robust decision-making tool to enhance urban safety and preparedness against emergencies.

  2. Estimating care and support needs of older people at the small-area level: The case of Athens

    Alexandros Alexiou, Angelos Mimis, Thomas Georgiadis

    • Panteion University of Social and Political Sciences
    Abstract
    Population ageing poses significant challenges for health and social care planning, particularly when policy decisions require granular, sub-national data that national surveys rarely provide. This study addresses the data gap by generating high-resolution, small-area estimates of care and support needs among older people in Athens. We draw on data from the Survey of Health, Ageing and Retirement in Europe (SHARE), alongside the decennial Census of the population to produce neighbourhood-level estimates of local need. Iterative Proportional Fitting (IPF) is employed to construct synthetic populations whose sociodemographic profiles are constrained to match census track records. We apply spatial microsimulation techniques in order to estimate neighbourhood-level values of health and care indicators, such as limitations in daily activities, out-of-pocket care expenses, and perceived physical and mental health. We combine indicators and construct a multi-dimensional index, describing care and support needs among local older populations. We map results and explore resulting patterns, illustrating the intra-urban variation presented. We further analyse results by exploring the relationships between the observed level of need and sociodemographic characteristics of neighbourhoods. This work carries important implications for policy, particular within local government contexts. Local decision-making requires access to granular data for effective policy planning, and the resulting heterogeneity across neighbourhoods underscores the limitations of aggregate-level planning approaches. High-resolution care need estimates can be leveraged for more efficient resource allocation, targeted service delivery, and inform overall local government strategies, enabling authorities to identify priority areas, and plan interventions accordingly.

  3. DigiFarm: Covering the digital training needs of agricultural advisors in Europe

    Athanasios Askitopoulos1, Dimitrios Bormpoudakis1, Spryos Theodoridis1, Ioannis Kontogiorgakis1, Stavros Papadopoulos2, Maria Prantsidou3, Alessandro Lachetti4, Marianna Hadjichristodoulou3, Thrasos Stylianou3, Eleni Loulli3, Christiana Papoutsa3, Diofantos Hadjimitsis5, Argyro Tsafara6, Dimitrios Argyropoulos6, Menelaos Stavrinides5, Charalampos Kontoes1

    • 1National Observatory of Athens
    • 2New Agriculture New Generation
    • 3Eratosthenes Centre of Excellence
    • 4Confagricoltura – General Confederation of Italian Agriculture
    • 5Eratosthenes Centre of Excellence / Cyprus University of Technology (CUT)
    • 6University College Dublin
    Abstract
    With the development and expansion of smart and/or digital farming, geoinformatics is becoming increasingly central to agriculture. Yet, the effective uptake of Earth Observation, spatial data, sensors and decision-support tools into everyday farm practices remains slow and uneven. DigiFarm, an Erasmus+ project, addresses this gap by developing an open online training programme targeted to agricultural advisors, designed to strengthen spatial intelligence, digital literacy and sustainability-oriented advisory capacity. Building on a previous needs assessment of agricultural advisors in four European countries (Cyprus, Greece, Italy and Ireland) and on an extensive review of 26 existing digital-agriculture training initiatives/ courses, we present the development of DigiFarm’s modular training scenarios through the Cultivate Academy e-learning platform. The programme is organised around four interconnected modules: remote sensing technologies and applications; tools and data for smart farming; green and soft skills; and business development for digital agriculture. These modules place GIS, satellite imagery, sensor data, farm-management systems and decision-support workflows within real advisory contexts. Our pedagogical approach combines self-paced online learning, scenario-based examples, interactive quizzes, certification and quality-assurance indicators, and is aligned with vocational and adult-learning principles. We place emphasis on translating spatial and environmental data into practical recommendations concerning crop monitoring, phenology, soil and water conditions, resource efficiency, climate adaptation, and farm-level decision making. A common thread linking the modules of the programme is the integration of ethics-and-society concerns. This addresses data governance, trust in recommendations, digital tool access inequalities, AI and transparency, and the social role of advisors as mediators between technological innovation and farmers’ situated knowledge. By embedding these issues within geoinformatics training, DigiFarm frames spatial intelligence as a socially accountable advisory practice and not merely as a technical competence. We argue that this type of capacity-building is essential for useful and responsible geospatial applications in European agriculture.

  4. Mapping Apartment Housing in the Eastern Mediterranean

    Athanasios V. Argyriou, Konstantina Kalfa, Angelos Hliaoutakis, Nikos Papadopoulos

    • Institute for Mediterranean Studies, Foundation for Research and Technology-Hellas, 74131 Rethymno, Greece
    Abstract
    This research is part of the ERC-funded project MCH-EsMed: How the Middle Class Housed Itself in the Eastern Mediterranean, based at IMS, FORTH, which explores how inhabitants, in key Eastern Mediterranean cities, located themselves in the building environment and constructed their ordinary apartment buildings. Most of these cities are dominated by common apartment buildings, not grand landmarks, indicating the type of residential structures most people actually live. Examining the emergence and formation of such buildings requires analyzing large amounts of data collected during long periods of time, since thousands of contractors, architects, developers and residents all were actively involved in shaping the urban landscape. This study highlights the methodological framework to study this process utilizing GIS and other geospatial approaches. Using untapped demolition and building permit data, a combination of digital tools (GIS, photogrammetry) and traditional tools (censuses, surveys, aerial photography) was employed. Through these methods we can visualize and analyze the spread of apartment buildings over time, identify the contractors and key actors involved in their construction, and trace the architects who signed these permits. By bringing all of this together, what once appeared as an anonymous collection of identical buildings starts to reveal identifiable patterns, processes, and agents. It is possible to identify who build what, where and when by answering the following so far unaddressed questions: how was the wealth of the housing market in these cities distributed? Were there monopolies? What were the main strategies of developers in these real estate markets? Only by leveraging, we suggest, large-scale data in an iterative process can be meaningfully to understand how the Eastern Mediterranean cities were produced.

  5. Unknown Ownership as a Spatial Indicator of Territorial Abandonment: Evidence from the Hellenic Cadastre

    Sevasti Chalkidou1, Konstantinos Tsiouplakis2, Angeliki Tsorlini2, Themistoklis Roustanis2

    • 1School of Rural and Surveying Engineering, Aristotle University of Thessaloniki
    • 2School of Rural and Surveying Engineering, Aristostle University of Thessaloniki
    Abstract
    Cadastral registration in Greece was designed to produce a definitive record of property rights. It also produced an unintended, geographically uneven map of properties for which no one submitted valid tenure documentation. These "unknown ownership" entries are not randomly distributed. Their concentration at the municipal and regional level tracks patterns of rural depopulation, inheritance fragmentation, and the long absence of formal property documentation in peripheral areas. This paper treats unknown ownership as a spatial proxy for territorial abandonment. The analysis is based on the spatial data infrastructure of the Hellenic Cadastre (Ktimatologio) to calculate unknown-ownership shares for each administrative unit and to map the resulting concentrations. These patterns are then overlaid with CORINE Land Cover data to examine whether unknown ownership clusters disproportionately occur on agricultural and forest land, where property values are lower and incentives for formal registration are weaker. Census population data are integrated to test the association between unknown-ownership rates and demographic decline. Sequence matters here. Urban areas were surveyed first; rural and semi-urban areas came last. Recently surveyed non-urban zones are therefore expected to exhibit higher unknown-ownership rates, partly as an artefact of the cadastral process itself rather than ownership conditions alone. The paper also examines the structural reasons behind non-registration. In many rural areas, property deeds simply do not exist. Inheritance chains remain unresolved across generations. Informal tenure persists. And a problem that receives too little attention: the chronic lack of geometric documentation means that owners may hold a valid deed, yet cannot locate their property on a map. Ownership without spatial identity, in cadastral terms, amounts to no ownership at all. The findings speak to land governance failures and the geography of institutional neglect across southern Europe.

  6. Mapping dissolved oxygen (DO) spatial distribution of coastal aquaculture environment based on remotely sensed data and machine learning algorithm

    Androniki Dimoudi1, Christos Domenikiotis1, Dimitris Vafidis1, Giorgos Mallinis2, Nikos Neofitou1

    • 1University of Thessaly
    • 2Aristotle University of Thessaloniki
    Abstract
    Monitoring dissolved oxygen (DO) in marine environments where aquaculture operates is crucial for ensuring aquaculture productivity and preserving the ecological status of the surrounding ecosystem. This study aims to present and evaluate a methodological approach for predicting DO in a semi-enclosed cove, where floating cage farming is carried out. It integrates in situ measurements and remotely sensed data to develop a machine learning regression model. The study area is located in the southeastern part of Pagasitikos Gulf (Eastern Mediterranean). Surface DO measurements were carried out over six seasons, between summer 2021 and spring 2024, recording 152 measurements in total, using a CTD (Seabird-19plus). Atmospherically corrected Multispectral Sentinel-2 MSI images were selected close to the sampling dates. Based on the coordinates received during sampling, pixel-to-point matching was performed, and remote sensing reflectance (Rrs) was extracted using SNAP (v.13.0). A selected number of bands from the satellite data was imperative for applying machine learning models to remove collinearity. For this purpose, two criteria were utilized: (a) the Variance Inflation Factor (VIF ≤ 10) index and (b) the correlation coefficient (rho) between spectral bands and DO, using IBM SPSS Statistics (v.31). The support vector regression (SVR) model was developed using the Python scikit-learn library with 5-fold cross-validation. SVR was selected as it demonstrates better performance compared to other regression-based machine learning (ML) models and neural network (NN) architectures under limited data availability. The SVR was trained using five of the six sampling campaigns. The remaining season was utilized for validation, comparing the actual in situ measurements with the extracted predicted values. The model performance was examined with different statistical indices. The resulting map shows the spatial distribution of DO across the entire study area. This approach could serve as a basis for the systematic monitoring of DO in coastal aquaculture environment.

  7. Spatial Analysis and Local AI for Smart City Digital Twins: A Framework for Real-Time Urban Risk Monitoring

    Dimitris Kavroudakis

    • University of the Aegean, Department of Geography
    Abstract
    Compound risks are growing in urban environments: infrastructure strain, crowd pressure, microclimate variability and operational disruptions. City planners need evidence-based, real-time tools to promote resilience and adaptive management. In this paper we propose a smart city framework that integrates IoT sensing, local (on-premise) AI for time series anomaly detection, spatial analytics, and Digital Twin models for enhanced urban monitoring and decision making. The framework assumes the deployment of IoT devices in urban spaces to capture variables such as temperature, humidity, CO2, vibration and population density. Data is sent in near real-time with lightweight protocols and stored locally in time-series and GIS-enabled spatial databases. A local-first architecture provides low latency, operational autonomy, and satisfies data governance and privacy requirements. We formulate an end-to-end anomaly detection pipeline that includes data ingestion, data validation, pre-processing, feature engineering, and model inference. It allows for different analytical approaches, from statistical baselines to machine learning and neural models (e.g. autoencoders, LSTM) to detect complex and non-linear patterns in urban dynamics. Anomaly outputs are integrated in a Digital Twin environment, combining 3D visualizations of buildings or districts with dynamic layers, such as environmental conditions, mobility flows and scenario simulations. Spatial AI methods (e.g., spatio-temporal clustering, hotspot detection) translate anomalies into actionable insights such as risk maps and vulnerability indices. In general, the framework allows smart city management to transition from reactive monitoring to predictive, interpretable decision support, allowing early warning systems, optimized urban operations, and better resilience planning.

  8. A Geospatial Data Integration Framework for Urban Mobility Analysis in Small Island Cities: The Case of Mytilene, Greece

    Nikoleta-Nefeli Kofou1, Dimitris Kavroudakis2

    • 1Laboratory of Spatial Analysis Geographic Information Systems and Remote Sensing (SAGISRS), Department of Geography, University of the Aegean, University Hill, 81100 Mytilene, Greece
    • 2University of the Aegean, Department of Geography
    Abstract
    Urban mobility analysis increasingly relies on the integration of heterogeneous geospatial datasets that describe the physical structure, functional characteristics, and population dynamics of cities. However, in many small island cities, mobility research is constrained by fragmented datasets, limited availability of traffic information, and the absence of integrated geospatial data infrastructures. These limitations hinder the systematic analysis of urban mobility patterns and the development of evidence-based planning strategies. This study proposes a geospatial data integration framework designed to support urban mobility analysis in small island cities. The framework combines multiple spatial datasets, including road network topology, pedestrian infrastructure, land use distribution, population patterns, and mobility-related observations, within a unified GIS environment. By harmonizing data from different sources and spatial resolutions, the proposed approach enables the creation of a comprehensive geospatial database suitable for advanced spatial analysis. The methodological workflow includes spatial data preprocessing, network structuring, attribute harmonization, and the integration of spatial indicators that describe accessibility, connectivity, and urban activity patterns. The framework is applied to the island city of Mytilene (Lesvos, Greece), a representative example of a small island urban system characterized by spatial constraints and heterogeneous mobility dynamics. By addressing challenges related to data fragmentation and interoperability, the study contributes to the development of geoinformatics methodologies for urban mobility research and demonstrates how integrated geospatial data infrastructures can enhance spatial intelligence applications in urban analysis and planning.

  9. Optimizing Spatial Decision-Making: Integrating Multi-Criteria Decision Analysis (MCDA) and GIS for Robust Site Selection

    Vasiliki Konstantinidou

    • National and Kapodistrian University of Athens
    Abstract
    Contemporary spatial planning faces the challenge of balancing economic goals with environmental and social needs. This paper presents an application that combines Geographic Information Systems (GIS) with Multi-Criteria Decision Analysis (MCDA) to provide a clear and scientific way to select the best possible sites for specific land uses. The methodology focuses on a screening stage that uses Boolean logic to apply "hard constraints" - such as legal restrictions and safety zones. This process effectively filters out unsuitable areas, leaving only the land that meets all necessary requirements. Within this framework, the criteria are evaluated to ensure the final decision is logical, consistent, and mathematically sound. This approach is demonstrated through a case study in Naxos, Greece. The analysis incorporates various parameters, including administrative boundaries, proximity to infrastructure, urban patterns, and environmental zones like NATURA 2000. Using official open data (from ELSTAT, GET SDI Portal, and Corine Land Cover), the entire workflow is automated within ArcGIS Pro using ModelBuilder to ensure the process is fast and easy to repeat. Ultimately, the results show that the combination of MCDA and GIS provides a reliable way for stakeholders to make objective, data-driven spatial decisions instead of relying on subjective judgment.

  10. Assessment of urban green spaces and development of a vegetation cooling index using PRISMA hyperspectral data

    Olympia Kourounioti

    • University of West Attica
    Abstract
    The management of urban ecosystems is a critical challenge in today's geographical research to ensure the quality of life in cities. Climate change and the Urban Heat Island (UHI) effect are changing the microclimate, making it essential to accurately map vegetation. This study focuses on Heraklion, Crete, a typical Mediterranean island city that experiences intense heat and persistent drought during the summer. This study aims to propose a methodology for assessing vegetation conditions and recommending targeted cooling measures. To this end, hyperspectral data from the PRISMA satellite were used for two dates (at the beginning and the end of summer). After preprocessing, the Normalized Difference Vegetation Index (NDVI) was applied to detect pixel density of vegetation, as well as the Normalized Difference Infrared Index (NDII) utilizing NIR (820nm) and SWIR (1650nm) bands, to assess the water status and health of the plants. Next, using the Linear Spectral Unmixing (LSU) method, density maps were created for five categories (ground, buildings, roads, water, vegetation). Finally, a cooling index was developed as the combination of vegetation density and vegetation health. The analysis revealed a significant deterioration in terms of health and a decline in the presence of vegetation during the summer months. In late summer, the number of pixels identified as “live vegetation” decreased significantly (~6.3 hectares), while those requiring immediate care increased, indicating severe dehydration. The study's findings highlight the fragility of urban green spaces under Mediterranean drought conditions. The resulting “Red Alert” map serves as a valuable decision-making tool for urban planners and stakeholders. As it is based on satellite data and vegetation indices, this methodology can be easily applied to any comparable urban center in the Mediterranean or Southern Europe facing similar climate challenges.

  11. Immersive Geovisualization of the Moria Refugee Camp in Virtual Enviroments

    Lemonia Moysidou, Ermioni-Eirini Papadopoulou, Stavros Proestakis, Nikolaos Soulakellis, Ioannis Chorianopoulos, Electra Petracou

    • University of the Aegean
    Abstract
    The aim of this study is to present the development of a digital twin of the Moria Refugee Camp on Lesvos Island, Greece, documenting its spatial transformation from 2017 to 2026 as an educational tool to enhance students understanding of spatial change in territories through geovisualization. Even though recent developments in geospatial technologies, digital twins, and immersive environments are available, there is still limited research concerning geovisualization methods for the effective representation of such complex spatiotemporal datasets. The Moria Refugee Camp is a suitable case for investigation, as it constitutes a dynamic case study characterized by continuous spatial transformations, shifting configurations, and changing land-use patterns over time. An innovative methodology was developed, integrating: 1. high-resolution UAV imagery and 360° panoramas, 2. image-based 3D modelling techniques, and 3. immersive visualization approaches. To represent the temporal evolution of the refugee camp, separate reconstructions corresponding to selected years between 2017 and 2026 were created and implemented within a game-engine environment. Due to restricted access during earlier operational phases of the camp, systematic field-based data collection was not feasible. To address this limitation, archival aerial imagery and journalistic visual material were additionally utilized to approximate the spatial organization and infrastructure density of earlier periods. The resulting application provides an immersive environment which allows users to navigate through the Moria Camp spatiotemporally. It allows the users to select different temporal stages in the evolution of the Moria camp and it can assist as a visual and conceptual aid towards the understanding of the concept and phenomenon of territorialization. Particularly, how state policies and practices organize space in order to control movements of people. Moreover, to assess the interaction of the above with the initiatives and actions of different organizations, asylum seekers and refugees in shaping the reality of space, power relations and human movements.

  12. A Methodological Framework for Assessing Tourism Hotspots at Fire and/or Flood Risk – Implementation in Four Greek Islands

    Angeliki Papazoglou, Anastasia Stratigea

    • Dept. of Geography and Regional Planning, School of Rural, Surveying and Geoinformatics Engineering, National Technical University of Athens
    Abstract
    In the Climate Change (CC) era, frequency and intensity of wildfire and flood occurrences are predicted to further amplify, resulting in severe socio-economic and environmental implications for the affected areas. Repercussions of such occurrences are expected to be more intense in the Mediterranean highly-urbanized, tourist-overloaded insular territories, highlighting the necessity for proactive –e.g. local/regional plans, land use etc.– but also reactive planning –e.g. Disaster Management Plans – in, mainly, tourism overloaded areas. In both proactive and reactive planning, efforts have to be carried out on the ground of the combined assessment/mapping of tourism hotspots and the spatially-defined fire/flood risk. In order for such a task to be accomplished, this work develops a framework for identifying tourism hotspot zones that are prone to fire and/or flood risk; and implements this framework in four Greek islands – Evia, Lesvos, Chios and Rhodes – as examples of different types of Mediterranean insular territories, exposed to both CC and (over)tourism. This framework integrates: i) a pool of fifteen literature-based, wildfire-related subcriteria that fall into four major criteria/factors, i.e. Topography, Climate, Environment and Urban Space, and are used as input to the Analytical Hierarchy Process (AHP) for producing, with the support of the Geographical Information Systems (GIS), Fire Risk maps for the study areas; ii) burned areas in the study islands in the time span 2000-2024 for validating fire risk assessments; iii) the officially-delineated flood-prone zones for identifying flood risk in the areas at hand; iv) tourist hotspots in the study regions, identified by means of tourist accommodation distribution. Results obtained by implementation of this framework demonstrate the urgent need for setting up proactive/reactive strategies/plans since coastal urban sprawl due to tourism pressure in the studied islands is steadily leading to the expansion of built up space in areas of high fire and/or flooding risk.

  13. Urban Informatics for Spatial Inequality: Making 3D Land Administration Accessible for Mapping Vertical Environmental Injustice

    Georgios Potamousis, Dionysia Perperidou

    • University of West Attica
    Abstract
    Keywords: 3D Land Administration; LADM Edition II; Vertical Inequality; Urban Informatics; Spatial Justice; Environmental Justice; Voxel-Based Models; Traditionally, geoinformatics has focused on horizontal analysis at the neighborhood level, often overlooking how environmental factors such as air pollution fluctuate between different building floors. This study introduces a new workflow that connects geographic theory with practical technology by shifting from broad area analysis to detailed, parcel level assessment. Using the updated ISO 19152:2024 standard (LADM Edition II), we present a 3D Land Administration framework that incorporates atmospheric data directly into property records. Our approach uses high resolution 3D models of city structures and voxel-based pollutant dispersion models to measure air quality within dense urban areas. A voxel is a three dimensional unit, much like a digital building block that allows for detailed mapping of pollution as it moves and accumulates in the air. These models simulate the flow of pollutants through each "block," helping planners understand which parts of buildings or floors are most affected. By integrating these environmental datasets with the Unique Immovable Property Number (KAEK), the official cadastral identifier, we identify systematic vertical inequalities. In Athens, Greece, a city of dense apartment blocks, residents on lower floors typically experience poorer air quality due to pollutants trapped at street level within urban canyons. The implementation of a 3D Cadastre by authorities facilitates the identification of affected floors and the subsequent implementation of targeted interventions, such as enhanced air filtration, with the objective of reducing exposure. This framework transforms the Cadastre from a static registry into a dynamic tool for Spatial Intelligence. This approach ultimately empowers stakeholders to identify and address hidden vertical disparities, thereby upholding the "Environmental Right to the City" and promoting fair urban planning for all.

  14. Towards Alternative Indicators of Touristification: An Approach based on GeoAI and ML

    Dimitris Psarologos, Athina Avagianou, Maria Papadimitriou

    • University of the Aegean
    Abstract
    This study presents a reproducible GeoAI-based framework for detecting and spatially analyzing swimming pools as proxy indicators of touristification in tourism-intensive island environments. The methodology combines pre-trained and custom-trained deep learning models optimized for small-object detection in high-resolution imagery, enabling automated, large-scale extraction of swimming pool features across Rhodes, Greece, for the period 2017–2024. Spatiotemporal analysis of the detected pools reveals a significant increase in their total number over the study period, with expansion extending well beyond established coastal resort zones into semi-rural and inland areas — a pattern that mirrors broader tourism and accommodation growth. To assess the touristic character of this expansion, detected pools are spatially associated with hotel and short-term rental locations; results confirm strong proximity, indicating that pool proliferation is predominantly tourism-driven. Further spatial analysis identifies considerable overlap between pool concentrations and active private groundwater extraction points, suggesting that tourism-induced landscape transformation is directly intensifying pressure on the island's already scarce freshwater resources. The study's principal contribution lies in documenting the intensification of touristification across Rhodes through the spatial diffusion of swimming pools, while addressing a methodological gap in Tourism Studies by demonstrating how GeoAI-derived datasets can serve as a foundation for spatial intelligence workflows that support environmental monitoring and the identification of touristification processes. Given the post-COVID rebound of tourism with limited sustainability transitions, and amid growing concerns over global freshwater depletion, the proposed framework offers a transferable approach applicable to other socio-spatial contexts facing analogous pressures.

  15. Integrating Disaster Event Records and Earth Observation Data: A Geospatial Database for Spatial Intelligence in Southern Europe

    Olga Roussou, Kiosa Irint, Eftychia Sorligka

    • University of the Aegean, Department of Geography
    Abstract
    A complex panorama of natural hazards, including wildfires, floods, heatwaves, earthquakes and landslides, is increasingly prevalent in Southern Europe. However, effective disaster management is often constrained by fragmented and heterogeneous datasets. Global databases such as EM-DAT provide essential standardized information on disaster types and impacts, yet they frequently exhibit significant temporal and spatial gaps, particularly for localized or small-scale catastrophes. To address these limitations, this study proposes a geoinformatics framework for the development of an extensive geospatial database of natural disasters in the Mediterranean region. The proposed methodology adopts a multi-source data fusion approach, integrating high-resolution satellite-derived datasets from NASA, harmonized infrastructure datasets, and official event records from EM-DAT. The study improves the geographical completeness of traditional records by utilizing Earth observation (EO) data, enabling more accurate identification of the extent and frequency of hazard events. A strict harmonization and integrated process carried out in a PostgreSQL/PostGIS environment, that guarantees consistency across many spatial reference systems, temporal resolutions, and danger classification methods, is essential to this strategy. The resulting geodatabase highlights the critical role of geoinformatics in standardizing disaster records and serves as a critical decision-support tool. It facilitates multi-hazard risk assessment, supports climate change adaptation strategies, and provides the spatial intelligence required for evidence-based urban planning and cross-border emergency management in Southern Europe. Overall, the study presents a scalable framework for improving disaster data integration, enhancing vulnerability analysis, and strengthening disaster resilience across the region.

  16. Time Series Analysis in Geographical Data: (Seasonality) and (Cyclicality)

    Nikolaos Sorkos1, Nikoleta-Nefeli Kofou1, Dimitris Kavroudakis2

    • 1Laboratory of Spatial Analysis Geographic Information Systems and Remote Sensing (SAGISRS), Department of Geography, University of the Aegean, University Hill, 81100 Mytilene, Greece
    • 2University of the Aegean, Department of Geography
    Abstract
    This study aims to analyze the time series of the Normalized Difference Vegetation Index (NDVI) using the technical analysis indicators MACD (Moving Average Convergence Divergence) and RSI (Relative Strength Index), with the objective of detecting, identifying, and interpreting changes in vegetation dynamics during the five-year period 2018–2022. The research was conducted in the selected study area through the utilization of high spatial and temporal resolution satellite data obtained from the Sentinel-2 platform, enabling the monitoring of vegetation dynamics at both seasonal and multi-annual temporal scales.The methodological approach was initially based on the calculation of mean monthly NDVI values for the study period, followed by the application of the MACD and RSI indicators in order to examine the temporal variability of vegetation. Specifically, the MACD indicator was employed to identify turning points and shifts in vegetation dynamics through the convergence and divergence between moving averages, while the RSI indicator was applied to capture the intensity and rate of change of NDVI values, allowing the distinction between periods of vegetation strengthening (increased greenness) and weakening (reduced greenness).The results demonstrate the effectiveness of both indicators in the temporal monitoring of vegetation and their ability to capture variations primarily associated with the seasonal and cyclical behavior of vegetation. The MACD proved to be more effective in identifying long-term trends and points of trend reversal, whereas the RSI showed greater sensitivity in detecting short-term fluctuations and distinguishing periods of relatively strong or weak vegetation activity. Overall, the findings indicate that technical analysis indicators such as MACD and RSI can provide a complementary and innovative approach to the analysis of vegetation time series. Their application to satellite-derived and geospatial data contributes to a better understanding of NDVI temporal patterns, particularly in relation to seasonality and cyclicality, and supports the interpretation of vegetation dynamics over time.

  17. Geospatial Analysis of the Energy Landscape in the European Union with emphasis on Electricity

    Konstantinos Statheropoulos, Marinos Kavouras, Athanasia Darra

    • National Technical University of Athens
    Abstract
    The ongoing energy crisis and the transition towards sustainable energy systems have highlighted the need for comprehensive spatial analysis of energy patterns across the European Union (EU). This study presents a geospatial approach to evaluating the energy landscape of EU member states, with a particular emphasis on electricity. The main objective of this research is to identify spatial patterns and disparities in energy production, consumption, and efficiency, and to propose a representative energy model that could guide future policy decisions. The study integrates multiple datasets from international energy organizations and statistical sources, incorporating key indicators such as energy intensity, electricity production by source, and the contribution of fossil fuels and renewable energy. A Geographic Information Systems (GIS)-based methodology is applied to process, analyze, and visualize the data. Spatial indicators are developed to enable comparative evaluation among EU countries. The results are presented through thematic maps and statistical visualizations, allowing for a clear interpretation of spatial trends and temporal variations. The analysis reveals significant differences in energy profiles across member states, reflecting diverse economic structures, policy frameworks, and levels of energy transition. By evaluating these differences, the study identifies best-performing countries and highlights potential pathways for improving energy efficiency and sustainability across the EU. The research contributes to the field of geoinformatics by demonstrating the value of spatial analysis in understanding complex energy systems. Furthermore, it provides a decision-support framework for policymakers, supporting the development of a more integrated and sustainable European energy strategy.

  18. A Spatiotemporal Pipeline for Detecting and Visualizing PM2.5 Pollution Events Using Low-Cost Sensors

    Sofia Zafeirelli1, Martin Dolejš2, Dimitris Kavroudakis3

    • 1University of the Aegean
    • 2Jan Evangelista Purkyně University in Ústí nad Labem
    • 3University of the Aegean, Department of Geography
    Abstract
    Urban growth has made air quality issues harder to overlook, especially in dense cities where emissions tend to build up rather than disperse. That shift seems to place new pressure on monitoring systems, not only to collect data but also to interpret it quickly enough to flag meaningful pollution episodes. This study focuses on Athens, Greece, using hourly PM2.5 data collected through a network of low-cost sensors. While these sensors are becoming more widespread, integrating their data into a consistent analytical workflow is still not straightforward. Despite the volume of available data, there is no widely used framework that combines cleaning, event detection, and spatial analysis at a fine temporal scale, which can limit how effectively the data is used. The approach developed here includes preprocessing, event detection, and spatiotemporal analysis. Between 2018 and 2022, more than 400 pollution events were identified. Many of these occurred during winter evenings and are likely linked to residential heating, along with atmospheric conditions such as temperature inversions. The framework does appear to capture large-scale events as well, including the wildfires in August 2021. Inverse Distance Weighting (IDW) was used to estimate pollution levels between sensor locations. Tracking the movement of event centroids provided a way to examine how these episodes expand, shift, or dissipate over time. While this method involves certain assumptions (for example, IDW may smooth local extremes), it offers a useful representation of spatial dynamics. Taken together, the results suggest that low-cost sensor networks, when handled carefully, can yield fairly detailed and reliable insights into urban air quality. They are perhaps best understood as a complementary layer, which adds granularity and responsiveness. With further refinement, the framework could be adapted for near real-time use, potentially informing local environmental management and public health responses.