Analysis Of Shared Bicycle Riding Based On Spark

The development of shared bicycles quickly provides a convenient way for people to travel short distances.It is one of the ways to realize “network traffic” and an innovative mode of urban short-distance traffic systems.It has the characteristics of convenience,low-carbon and environmental protection,and effectively solves the “last mile” problem when people travel.It also affects the daily mode of travel of urban residents and slowly changes the spatial pattern of the city.However,with the rapid development of the bicycle sharing industry,it has also brought about extensive operations such as serious parking disorder and unbalanced resource allocation.How to fine-tune the operation of shared bicycles and realize dynamic resource allocation has become an important issue in the development of shared bicycles.This paper uses the Python web crawler to obtain Mobike data from Tianjin and combines the time and space characteristics of riding behaviors with the Apache Spark big data platform and visualizes the results.The research content of this paper is as follows:(1)Built a large data processing platform based on big data technology to analyze the use of shared bicycles with Spark,and demonstrated the flexibility of data through the personal version of the BDP.(2)Using the Python web crawler to obtain datasets of Mobike in Tianjin,and using Python and Spark SQL as programming languages to perform data preprocessing on the crawler data and storing the data cleaned on HDFS.(3)Using Spark SQL to analyze the total Mobike cycles and Mobike cycle utilization rate in Tianjin urban area in August;and to analyze and analyze Mobike cycling rides from Mobike cycling distance,riding frequency and heat map respectively;Based on the bicycle cycling situation at different time periods on weekdays and rest days,the main reasons for forming the travel time characteristics are analyzed.(4)Modeling the geographic location of the Mobike and their return by the clustering algorithm in the Spark MLlib,and evaluating the clustering model;combining the clustering prediction results with the analysis of the starting position The demand characteristics of bicycles at various clustering points also analyze the demand characteristics of borrowing and retrieving vehicles at each cluster point in the morning and evening peak periods,and propose relevant suggestions for the demand characteristics.