A statistical method of population spatial distribution based on regional grid and K-means clustering algorithm

Abstract

With the progress and development of science and technology, mobile phone signaling data and new statistical methods bring new research methods and directions for population statistics. Regional grid algorithm is an algorithm that further refines the sector, divides the sector into a single and equivalent grid, counts the population in each grid, and further improves the statistical accuracy of population density and distribution. K-means clustering algorithm adopts Euclidean distance as the evaluation index of similarity, which is a typical distance-based clustering algorithm. Based on the instant call record data of mobile communication operators, using the location, time, personal attributes and other information in the mobile signaling data, combined with the base station engineering parameters, and based on the demographic behavior characteristics, this paper judges and measures the mobility of the population from the user behavior represented by mobile communication big data. Regional grid algorithm and K-means clustering algorithm are used to process mobile communication data, and the results are classified, so as to analyze the permanent population size of each region in Wuhan. Through case analysis and verification, the error of classification results is far less than 0.05, and the classification is basically in line with the population distribution of each region in Wuhan, which can verify the validity of the results.