Graph-Laplacian modeling gives more reliable house price estimation, research shows

Marc Francke and Willem Sijp published a new peer-reviewed article in Real Estate Economics¹: “Graph-Laplacian modeling of spatiotemporal effects for house price estimation”. The open access paper explores how graph-based methods can improve house price estimation and local price indices. Marc’s affiliation with both the University of Amsterdam and Ortec Finance underlines the link between academic research and practical valuation expertise.

The research addresses a familiar challenge in real estate markets: local transaction data can be sparse, especially at neighborhood or monthly level. Standard models may become noisy when there are few comparable sales. The Graph-Laplacian framework helps by sharing information across neighboring locations and adjacent time periods in a structured way.

Key benefits of Graph-Laplacian modeling

A key benefit is that the model can produce more stable estimates in thin markets, where transaction volumes are limited. By capturing relationships between nearby properties, regions and time periods, the method helps avoid overly volatile local estimates. This supports more reliable property-level prediction and granular price index construction.

Methodologies compared, and its application

The graph-Laplacian can be input for a prior in a Bayesian estimation setup or used as regularization term in a Ridge regression. A spectral decomposition of the graph-Laplacian significantly reduces computation time for estimation.

As an application, the authors estimate graph-Laplacian hedonic pricing and repeat-sales models on sales prices of Australian residential properties in the period from 1990 to 2024. Bayesian and Ridge regression estimation results are very similar, although the computation time for the Ridge regression is orders of magnitude faster, however at the expense of missing posterior density functions.

For Ortec Finance, the publication is relevant to its broader work in real estate valuation, market modeling and property-tax related analytics. Better granular price models can help strengthen the analytical foundation behind fair, data-driven property valuation processes.

Read the article here

For more information about the research and its practical implications, please contact Marc Francke.

¹Sijp, W. P., & Francke, M. K. (2026). Graph-Laplacian modeling of spatiotemporal effects for house price estimation. Real Estate Economics, 1–24.