Housing and the social domain

Population ageing is not just a demographic trend; it has a clear spatial pattern. It concentrates in particular neighbourhoods, streets and building types, and so does the gap between the homes people live in and the care and support they need or may need in the future.

Knowing averages is not enough: meaningful decisions require understanding where needs are concentrated, in which neighbourhoods, streets, blocks and types of housing.

We bring more than 25 years of experience in housing, care and the social domain, combining insights from applied gerontology with expertise in spatial analysis and modelling. We support municipalities, care providers and housing corporations with questions about suitable housing, the location and accessibility of services, and the planning of the living environment. Privacy is an important consideration for us and therefore a key design principle: our models transform personal information into spatial insights, enabling societal use while protecting individuals.

WoonZorgwijzer

The WoonZorgwijzer is our most widely used and adopted application in the social domain. It has been adopted by municipalities, housing corporations, healthcare providers, and many other organisations.

The WoonZorgwijzer provides detailed spatial estimates of the prevalence of health conditions and the functional limitations associated with them. Health conditions support strategic housing and care policy, while functional limitations help organisations identify where specific accessibility or support measures are likely to be needed. Together, they connect population health with the practical implications for housing and care.

The WoonZorgwijzer estimates the prevalence of health conditions from demographic and household characteristics. It does not use information about the health conditions or functional limitations of identifiable individuals. The resulting spatial patterns provide relevant insights for developing housing and care policies and programmes at different levels.

The tool is developed and maintained through collaboration between several organisations. Stichting In Fact develops the models that estimate the prevalence of health conditions and functional limitations mainly based on CBS microdata and makes these estimates available as datasets at an enhanced postcode level. Our role is in the geographic modelling: transforming these estimates into spatial representations and presenting them through interactive maps. BaasGeo develops the viewer together with us. A broader partnership, including Platform31, the Ministry of the Interior, several provinces and municipalities, housing corporations and RIGO, contributed domain expertise, policy insights, and practical knowledge to shape the instrument.

See web-based model visualisation for how it is built and who uses it, and the project list for the provinces, regions and municipalities it runs for.

Plus variants

The WoonZorgwijzer Plus combines the WoonZorgwijzer with additional layers covering topics such as:

  • demographic projections;
  • socio-economic indicators;
  • the physical suitability of the housing stock;
  • the location of services;
  • walking distances to services;
  • data on housing corporation supply.

In these applications, the WoonZorgwijzer layers provide the foundation for understanding the demand for suitable housing, well-being, and care services. The additional layers are mainly used to provide insight into the supply side. By spatially confronting demand and supply, the WoonZorgwijzer Plus provides actionable insights for policy and planning.

How it is meant to be used

The WoonZorgwijzer and its Plus variants are not intended to prescribe solutions or make decisions on behalf of users. Instead, their main purpose is to serve as a communication tool that facilitates collaboration between different stakeholders by providing shared, location-based insights.

An illustrative example of the use of the tool is presented in this Dutch video.

Modelling techniques

In the housing and social domain, we use a range of modelling techniques. Three important examples are:

Spatial pattern analysis and visualisation

Administrative boundaries are commonly used to present spatial data, for example on municipality maps. However, relying on these boundaries has disadvantages:

  • the boundaries themselves strongly influence the visual interpretation of the data;
  • spatial variation within administrative units is hidden;
  • spatial patterns that extend across administrative boundaries may remain unnoticed, and abrupt differences at their borders may be exaggerated.

To reveal the underlying spatial patterns, we often prefer to use focal statistics to smooth and visualise the data independently of administrative boundaries. The domain data itself should tell the story.

Accessibility modelling

The mobility of people is often influenced by the limitations associated with their age, health or social circumstances. As a result, the accessibility of essential services becomes an important factor in assessing the suitability of locations.

In the real world, modelling accessibility using straight-line ("as-the-crow-flies") distances often produces unrealistic results. Roads, footpaths, waterways and other barriers determine how people actually travel. Therefore, we use network analysis (see also networks and accessibility) to calculate realistic travel distances and times. For example, we can determine the network distance from every (suitable) dwelling to the nearest supermarket.

Local-scale forecasting

Decision-making in the social domain often involves real estate and public facilities. These decisions are location-specific and typically have a long-term perspective. Therefore, we estimate not only the current demand, but also the future demand at a local scale. While such projections inevitably involve uncertainty, they provide valuable insights and form a sound basis for discussion and decision-making.

Data used

In this domain, we often combine domain-specific datasets with national registers, including addresses and buildings (BAG), topographic data (BRT and BGT), elevation data (AHN), cadastral information (BRK), and property valuations (WOZ). A key strength of our work is the ability to effectively integrate these.

Domain-specific data is collected from a wide range of sources, including public websites, open datasets, and information provided by local/regional organisations. Collecting, validating, and harmonising these datasets is often a substantial part of a project. Local and regional knowledge is an important quality factor, as the quality and completeness of nationally available domain-specific data are often limited.

Protecting privacy is a core principle of our work. Wherever possible, our models rely on aggregated or anonymised information rather than personal data. When the use of personal data is unavoidable, we use only what is strictly necessary, never disclose personal information about individuals, and process all data in accordance with our privacy protocol (in Dutch).

With regard to the national registers, making these datasets usable for spatial modelling is a substantial part of our work. We develop and publish the tooling required to process and use these sources, including our BAG tools and Pand Hoogte Nederland, which derives a height for every building in the country from the national elevation rasters.