Research on big data mining of traffic flow based on GIS and Python

MENG Lihua; WANG Shiguang

doi:10.19301/j.cnki.zncs.2025.05.006

Abstract：The time-varying characteristics of urban road traffic volume is an important indicator to characterize urban traffic conditions.The study proposes a detector data processing and analysis method based on GIS and Python to study the time-varying correlation of multi-scale traffic flow.First,the vector road network is acquired based on GIS,the bayonet detector is matched to the vector road network,and the detector data preprocessing and quality evaluation are performed.Then,according to the meaning of each field in the data,a Python-based bayonet data processing process is designed,and the time-sharing traffic is obtained by statistics.Finally,based on the data of the bayonet detector in Haikou,the time-varying correlation and migration of traffic in multiple cities are analyzed.Studies have shown that there is a relatively stable linear relationship between the peak hourly traffic volume,hourly volume,and daily traffic volume in Haikou City.And between 8:00-19:00 on the same date,the linear fitting parameter values of hourly traffic volume and peak hourly volume are similar.Migration experiments in multiple cities have verified the applicability of the empirical model of Haikou data.The data fitting error between 7:00-18:00 is generally less than 20%.