Enhancing population data granularity: A comprehensive approach using LiDAR, POI, and quadratic programming

Abstract

This research presents a sophisticated framework for the precise downscaling of population data from census blocks to individual residential units, employing an integration of housing unit characteristics. The aim was to devise and substantiate a thorough methodology for the distribution of households within specific residential buildings. Utilizing the Microsoft Building Footprint dataset, LiDAR remote sensing, and Point of Interest (POI) data, a detailed inventory of residential structures was compiled. A quadratic programming model and Monte Carlo Simulation techniques were applied independently for the strategic allocation of households to these buildings. For validation, this study conducted a comparative analysis between the two methods. The outcomes revealed that the quadratic programming model provided superior precision and detail in population data compared to the Monte Carlo Simulation technique. Consequently, the quadratic programming model significantly enhances the granularity of population distribution data, offering a valuable tool for more informed decision-making.