Revisiting Urban Immovable Property Valuation: An Appraisal of Spatial Heterogeneities in Punjab Using Big Data

Abstract

This study employed big data and spatial analysis to assess property values in two cities, Lahore and Faisalabad. Traditional housing price models overlook spatial nuances, focusing solely on structural attributes. To address this, we constructed valuation models using ordinary least square regression and Fast Geographic Weighted Regression (FastGWR), implemented through Python and MPI, based on spatial variables. The models explained up to 75 percent of variance in Faisalabad and around 85 percent in Lahore. Factors like floor area, proximity of health facilities, recreational sites, and marketplaces add a premium to prices, while the nearness of educational institutions, worship places, and solid waste transfer stations or dumping sites lessen the property values in both cities. However, the proximity of industrial units and graveyards affects property values negatively in Lahore but positively in Faisalabad. This study highlights the critical significance of spatial factors in urban immovable property appraisal. As a result, it is recommended to integrate these factors into the process of policy formulation and urban planning.