Parcel-Level Land Valuation under Planning Policy Change: A Spatially Based and Segmented Modeling Framework

Abstract

Planning policies shape land markets by defining development rights, regulating land use intensity, and signaling future growth expectations. However, in current practice, land and parcel value estimation remains highly dependent on expert judgment and manual comparison of recent transactions. While professional appraisal methods are effective for individual assessments, they rely heavily on subjective interpretation and lack systematic mechanisms to identify how planning policy changes influence parcel values across large datasets. This limitation is particularly evident in growing suburban municipalities, where frequent policy updates are used to respond to development pressure and evolving growth objectives. As a result, both municipal decision-making and real estate analysis increasingly require data-driven models capable of automatically evaluating the valuation impacts of planning policy changes. This thesis addresses this gap by developing a spatially explicit, data-driven framework to examine how planning policy changes are capitalized into parcel-level land values in the Town of Aurora, Ontario. The research integrates GIS-based spatial analysis with segmented regression and machine learning modeling to move beyond manual valuation approaches. Parcel-level datasets are constructed by combining transaction records with Official Plan designations, zoning regulations, and adjacent based spatial variables. Parcels are further segmented by size into two datasets to distinguish between house driven and land driven valuation mechanisms, enabling the model to identify the conditions under which planning signals emerge in observed prices.