Socio-Spatial Analysis of Inequality with SoRa-Service: A Geolinking Approach

Socio-Spatial Analysis of Inequality with SoRa-Service: A Geolinking Approach Theodor Rieche, Denise Ehrhardt, Alexander Jung, Sebastian Eichhorn, Stefan Jünger, Benjamin Zapilko, Jan Goebel, Sujit Kumar Sikder, Gotthard Meinel Image credit: D. Ehrhardt Are residents of single-family houses really happier than those living in high-density housing? →Question addresses housing equity →Such and similar questions often cannot be answered alone with available statistical data Intro 07.11.2025 ILUS - International Land Use Symposium 2025 2 Source: D. Ehrhardt Compare: Odermatt, R., & Stutzer, A. (2022). Does the Dream of Home Ownership Rest Upon Biased Beliefs? A Test Based on Predicted and Realized Life Satisfaction. Journal of Happiness Studies. https://doi.org/10.1007/s10902-022-00571-w ▪Research questions ▪Data linking with Geolinking Service SoRa ▪Data ▪SOEP Survey Data ▪Urban Structure Type Data ▪Methodology ▪Results ▪Conclusion ▪Outlook Agenda 07.11.2025 ILUS - International Land Use Symposium 2025 3 ▪RQ1: How can survey data from the Socio-Economic Panel (SOEP) be linked to small-scale Urban Structure Type data and what potential does it have for socio-spatial equity research in the context of housing equity? ▪RQ2: What characterises the residents of different types of urban structures in Hesse? (An exploratory approach) Research questions 07.11.2025 ILUS - International Land Use Symposium 2025 4 ▪Linking survey and spatial data for socio-spatial research ▪User interface: a “sora” package in R language ▪Offers pre-defined linking methods and datasets ▪Private or public mode (inside or outside a secure room of RDC) ▪Privacy-compliant & FAIR principles ▪Currently SOEP and IOER Monitor data →extendable to other RDCs ▪Offering synthetic survey data to prepare the linking outside of RDC New Geolinking Service SoRa 07.11.2025 ILUS - International Land Use Symposium 2025 5 Key facts New Geolinking Service SoRa 07.11.2025 ILUS - International Land Use Symposium 2025 6 General workflow of data linking Spatial feature Survey variable ▪Annual surveys (“waves”) since 1984 ▪Representative sample of Germany (sample size: about 20.000 households with about 30.000 individuals) ▪Limited access for spatial reference of survey participants to avoid re-identification ▪Real address coordinates usable only on-site ▪Federal state level usable with data use agreement ▪Synthetic SOEP Structural Dataset for testing Dataset 1: SOEP Survey Data 07.11.2025 ILUS - International Land Use Symposium 2025 7 Socio-Economic Panel (SOEP) ▪Useful documentation: ▪SOEP Companion or paneldata.org Goebel, Jan et al. (2018). The German Socio-Economic Panel (SOEP). Jahrbücher für Nationalökonomie und Statistik. 239. 10.1515/jbnst-2018-0022. © DIW Berlin/Sandra Bohmann ▪100 m gridded data with Urban Structure Type (SST) ▪Spatial extent: Hesse (Germany) ▪Timestamp: 2022 ▪Data collection method: ▪AI-based classification (XGBoost) ▪Based on official building data (LoD2), cadastral parcels and remote sensing data ▪Overall accuracy: 95 % Dataset 2: Urban Structure Types in Hesse (SST) 07.11.2025 ILUS - International Land Use Symposium 2025 8 Keller, Sina et al. (2025). Entwicklung von Planungshilfen für Klimaschutz und Klimaanpassung in der räumlichen Gesamtplanung mittels Fernerkundung - Abschlussbericht. Hessisches Ministerium für Wirtschaft, Energie, Verkehr, Wohnen und ländlichen Raum. 10.5445/IR/1000182412 High-density buildings Multi-family houses High-rise buildings Row apartment buildings Outbuildings Low-rise buildings on large area Single-family houses Perimeter block ▪Explorative data analysis with 24 survey variables ▪Spatial aggregation based on household level ▪Including a weighting factor to improve representativeness ▪Simple point-to-raster linking (method: “lookup”) Methodology 07.11.2025 ILUS - International Land Use Symposium 2025 9 Workflow Results 07.11.2025 ILUS - International Land Use Symposium 2025 16 Nodata also includes owners who were not asked this question. Results 07.11.2025 ILUS - International Land Use Symposium 2025 17 High-density buildings 7,67 Multi-family houses 7,84 Row apartment buildings 7,30 Perimeter block 7,50 Single-family houses 8,07 Big city 7,59 Medium-sized town 7,82 Larger small town 8,07 Small town 8,10 Rural municipality 7,78 ▪The SOEP households were mostly assigned to residential building types and are well distributed across municipality types (except rural municipalities). ▪Dwellings with many rooms are more likely to be found in singleand multi-family houses, as well as outside of big cities. ▪Residents of single-family houses, small towns and rural municipalities change their homes less frequently. ▪Net household income is highest in singleand multi-family houses, as well as in rural municipalities. ▪Social housing is often located in row apartment buildings in big cities. ▪Satisfaction with dwelling is generally quite high and relatively consistent (with slight peaks for single-family houses, small towns and larger small towns). Conclusion 07.11.2025 ILUS - International Land Use Symposium 2025 18 RQ2 ▪Timestamps of selected datasets (2022 vs. 2020) ▪Sample Size and statistical confidentiality (e.g. spatial aggregation or share of nodata) ▪Aggregation between SOEP households and SOEP individuals ▪Weighting factor for households and individuals in survey data ▪Positional accuracy in 100 m grid cell (building footprint vs. address) & partly missing coverage of addresses ▪Uncertainties in classification of Urban Structures Types Conclusion 07.11.2025 ILUS - International Land Use Symposium 2025 19 RQ2 - Limitations ▪Geolinking Service SoRa was used to link an external spatial dataset (at IOER Dresden) to highly protected SOEP survey data ▪Demonstrated, how SOEP data and small-scale Urban Structure Types can be linked ▪Spatial typologies can help to aggregate survey data for the required purpose ▪SOEP Survey Data covers various topics & offers real address coordinates ▪SOEP Survey Data are highly protected to avoid re-identification of survey participants ▪Linkage can be prepared from outside the SOEP secure room using synthetic data, to save working time on-site Conclusion 07.11.2025 ILUS - International Land Use Symposium 2025 20 RQ1 ▪Various research questions can now be answered with the Geolinking Service SoRa ▪For example in the field of housing/environment/mobility equity ▪Using survey data from whole Germany could be useful (sample size) ▪Linking further spatial datasets, e.g. based on density or accessibility indicators ▪Include neighbourhood of survey participants (circle, isochrones etc.) Test period with Geolinking Service SoRa starts in January 2026 Are you interested to enrich your spatial model with individual perspectives, behaviour, or perceptions? Outlook 07.11.2025 ILUS - International Land Use Symposium 2025 21 www.ioer.de/en Thank you for your attention! Theodor Rieche [email protected] www.ioer.de Urban Structure Types in Marburg / Hesse, Google Maps Geolinking Service SoRa [email protected] www.sora-service.org Are you interested in linking your spatial science models with social science survey data?→Test period starts in January 2026 ! Results 23 Results 24 High-density buildings 1969 Multi-family houses 1953 Row apartment buildings 1965 Perimeter block 1945 Single-family houses 1968 Big city 1960 Medium-sized town 1972 Larger small town 1964 Small town 1963 Rural municipality 1958 Results 25 High-density buildings 85,93 Multi-family houses 114,45 Row apartment buildings 63,78 Perimeter block 71,17 Single-family houses 114,45 Big city 84,93 Medium-sized town 102,92 Larger small town 106,62 Small town 124,54 Rural municipality 135,68 Results 32 High-density buildings 7,14 Multi-family houses 7,29 Row apartment buildings 7,52 Perimeter block 7,51 Single-family houses 7,23 Big city 7,24 Medium-sized town 7,17 Larger small town 7,43 Small town 7,20 Rural municipality 7,33 Results 33 Results 34 Results 35 Results 36 Results 37 Nodata also includes owners who were not asked this question. Results 38