Conference Agenda

Housing vacancy is a complex and context-specific phenomenon, with significant implications for land use efficiency. In Western European countries, the reasons for vacancy are often explained by the owners' behaviour. Yet, we hypothesize that owners' decisions are influenced by the attractiveness of their property on the market and that the characteristics of housing vacancy may provide insight into fine-scale market dynamics. Most previous research on vacant housing characteristics has used data aggregated at the level of neighbourhoods, cities or regions. There has been few research based on parcel data and - to our knowledge - none for Western Europe, probably due to the lack of data.

This study combines water and electricity consumption data with field verification to identify vacant single-family housing and examine their characteristics. The Liège agglomeration was chosen as a case study due to its complex urban dynamics, including shrinkage due to industrial decline, suburbanization, and the increasing commodification of housing – trends that are widespread in Western Europe since World War II and that influences the housing market.

Logistic regression models were used to analyze the factors influencing housing vacancy, comparing variables related to the neighborhood (e.g., socioeconomic precarity, age distribution, changes in the number of housing units) and housing characteristics (e.g., property age, presence of a garage, cadastral income). Our analysis reveals distinct trends: areas with higher vacancy rates are characterized by substandard, low-value housing, while low vacancy areas show vacancy patterns related to personal convenience and comfort preferences. In both cases, factors at the housing level have more influence than the factors at the neighborhood level.

A simplified model, using only land and building register data, was then applied to all single-family housing parcels in the Liège agglomeration to identify the least attractive housing whose characteristics make them prone to vacancy. Hotspot analysis using the Getis-Ord Gi* statistic highlights concentration of vacant-prone housing and enables us to identify some patterns (enclosed areas, dead-end streets, old et narrow streets in village centres, etc.).

While actions to tackle vacancy are generally taken on a case-by-case basis and are very time-consuming, this study lays the foundations for thinking about urban strategies at the neighbourhood or street level, such as de-densification or street-wide renovation. By targeting housing that is prone to vacancy, this approach also addresses related issues, such as managing substandard housing and improving the overall attractiveness of urban spaces.

Urbanization and internal migration are interlinked and shaped by historical, economic, and administrative factors in India. Internal migration in India has emerged as a significant driver of demographic, spatial, and socio-economic transformation in urban centres. The city of Hyderabad, a major metropolitan hub in southern India, has witnessed a consistent influx of internal migrants over recent decades. Hyderabad is the capital of the state of Telangana in India. These migrants, varying widely in terms of their socio-economic backgrounds, regions of origin, and migration motivations, engage with the city in spatially uneven ways. Their residential choices—whether voluntary or constrained—reflect broader structural dynamics, local housing market configurations, and often, invisible forms of social exclusion. In urban India, residential segregation is prominently shaped by caste-based discrimination, socioeconomic disparities, and state-led planning practices that systematically marginalize certain groups. The aim of this paper is to investigate the spatial residential segregation and quantify the degree of segregation experienced by internal migrants in Hyderabad using a combination of spatial and statistical methodologies.

The research adopts a mixed-methods approach that integrates spatial analysis, quantitative data, and statistical tools. The empirical data includes primary data collected through fieldwork conducted in three wards in GHMC (Greater Hyderabad Municipal Corporation) that house significant migrant populations. GHMC is the biggest Municipal Corporation of Hyderabad. The study uses household-level data to extract key indicators such as migrant status, income and education level, occupation type, duration of stay, housing typology, residential preferences, etc., and spatial data to extract the residential typologies. At the spatial level, Geographic Information Systems (GIS) are employed to map the distribution of residential typologies. A total of 10 distinct residential typologies were identified based on built form, affordability, and resident composition. Each typology exhibits different levels of spatial integration and access to infrastructure. Informal settlements and peripheral colonies are most marked by segregation, both in terms of migrant concentration and infrastructural deficit, reinforcing a double marginalization—social and spatial.

For spatial analysis of residential segregation, Massey and Denton's (1988) framework is applied to examine key dimensions such as evenness, exposure, clustering, and concentration. To delve deeper into the factors influencing housing preferences, a Principal Component Analysis (PCA) was conducted. These findings were then validated using ANOVA to detect significant differences in housing preference across income groups. A significant statistical relationship is observed between the income level and the type of residential typology migrants prefer.

In conclusion, the research finds that internal migrants in Hyderabad experience varying levels of residential segregation shaped by socio-economic factors such as income, marital status, occupation, family size, safety, etc. The spatial-statistical approach employed here underscores the layered complexity of migrant urbanism, where migrants make active choices about where to live, and these choices are reflections of structural constraints.

Densification (Nachverdichtung) is a key urban strategy aimed at accommodating development needs while limiting the consumption of undeveloped land. By concentrating housing and services within existing urban areas, cities can reduce infrastructure costs, minimize travel distances, retain tax revenue, and strengthen social cohesion. Among the various approaches to densification, one crucial distinction is between developments that occur within existing legal planning constraints and those requiring changes to zoning regulations. The presented work focuses on the former.

In collaboration with a municipal land development agency (as practice partner), the authors, as a research institution, are developing the AGORA tool (Analytics for Ground Property and Real Estate Assessment). The tool supports strategic land management by identifying parcels in Hamburg that exhibit underused development capacity within the constraints of existing planning regulations.

To assess such densification potential, the project compares actual land use data from the ALKIS cadastral system with regulatory data from the XPlanung standard. As a first analytical step, the focus is on the Grundflächenzahl (GRZ), a planning parameter indicating the allowable ratio of building footprint to land area. A core challenge lies in aligning buildings to parcels and assigning parcels to planning areas with defined GRZ values. The method addresses these spatial mismatches and calculates build-out intensity on a parcel level.

Now in its third iteration, the comparison method is being tested within LIG and integrated into internal workflows of the AGORA tool. This case study demonstrates how available geodata can be used to evaluate densification potential within existing legal limits, providing a transferable model for data-informed urban land policy and sustainable development planning.

In this talk, we present the methodological foundation of the comparison algorithm and its iterative devolvement, shaped by the motivations of the involved partners and computational limitations associated to the data structure and the tools used. We share as well, the current results of the implementation, in relation to the user test done with our partner.

Could maps help our societies discuss land changes the same way topographic maps have been adopted for ages to locate entities and discuss routes? If such maps exist, could they improve our individual and collective capacities to understand what desirable densification is and how to make it happen? The stakes of densification call for more comparative studies, multidisciplinary research, more trust between public and private sectors, and more democratic debate. Current maps to discuss densification are grounded on indicators and abstract areas. Not everyone can interpret them. Their granularity is often too coarse to account for local variations in the contrasting impacts of denser urban environments.

This motivates our research. We study the production of a new kind of maps fitted to open discussions on densification. Buildings and how they change over time are good candidates to put on such maps. They are commonsense entities. Overlaying them with context like green field, street network, public transports help observing the increase of habitat in built-up areas, the saving or consumption of green fields, the development of new infrastructures, the development of low-quality housing programs with insufficient amenities. A model has been proposed to derive automatically maps of building changes on city regions using different topographic buildings surveys available on the past decades in Germany, the UK and France. These initial maps are further used to refine the highly generic Building Change model, which was limited to changes that could be detected between topographic features, identify ambiguous words and implicit hypotheses.

This clarification of shared words to discuss about building changes is grounded on a collaborative, multi-lingual and multi-perspective collection of building changes data, that embraces different context, Germany, the UK and France, which timespan is included in the past decades. Items can be either a change of a specific entity, or a change of a neighbour, or a change occurring at the level of the urban sprawl. Each item is associated with data (buildings surveys differentials, digital elevation models, properties, aerial imagery), visuals that can be displayed on a screen, and narratives. Narratives are acquired from local portals where municipalities describe their development projects and from sessions where the collection is presented to experts who comment on visuals and relate them to densification. The collection is produced manually using data that are produced automatically like differentials of vector data or differentials of digital elevation models.

A key aspect we explain more in details is to attach context to building changes items and to narratives. This is crucial for our capacity to generalise concepts and identify ambiguities. It is achieved by a formal model to store the provenance of items on the one hand and the provenance of narratives on the other hand. The provenance of an item is a full description of how the item was produced, including how the focus area was selected. The provenance of a narrative includes the state of the collection that was presented during the session and the order in which items were presented.