A User-guided Cleaning And Visual Analysis Approach For Traffic Positioning Data

The development of modern science and technology has greatly promoted the rapid progress in China's transportation sector.Taxi,cars,buses and other means of transports have become more and more widespread.The growing trend of motor vehicles has continued to grow rapidly,so the data obtained from motor vehicles is also increasing.With the increasement,massive data has brought great benefits,but due to its negative factors,such as data errors,it has brought difficulties to data analysis.In order to analyze traffic data more effectively and accurately,this article proposes a user-guided data cleaning visualization system.This article proposes a data cleansing visualization system based on B/S structure.It combines data cleaning and visualization techniques to analyze floating car data.Different data cleaning rules are given for different errors in traffic data.This system allows users to independently select feature vectors for data cleaning and visual display.The main work and achievements of this article are as follows:(1)Handle wrong data in mass data by data cleaning.Due to the existence of multiple problems such as diversification of data collection equipment,equipment malfunction,transmission network and environmental factors,the original data uploaded to the server exists diversified formats,no uniform standards,data loss and even data errors.The data that exists above situations is called "dirty data." If these data are not processed,it will seriously affect the visualization results and cause fatal errors in data analysis.In this regard,we will formulate effective data cleaning rules to deal with traffic dirty data,and data cleaning will better promote the accuracy of data visualization and in order to obtain effective and correct visualization results.(2)For the difficulty of extracting valuable information from massive data,such as when processing higher-dimensional data,many traditional data analysis methods are no longer applicable.In order to present these data more intuitively to analysts,this article will adopt data visualization and visual analysis technology.Visualization can express abstract data in a more intuitive way for users to observe and analyze,improving the efficiency and accuracy of analysts;Moreover,this system satisfies the user's own needs.Because different users have different requirements for the visualization of traffic data,the system allows users to customize data cleaning and visualization of feature vectors based on their own needs.This can improve the system's applicability and wide use;Since traffic data basically exists in units of point,it is difficult to effectively evaluate and display the cleaning effect before and after data cleaning.This article uses an autonomously designed map view.We proposed a boundary-resevered map deformation approach for visualizing geographical map which is called a “Rectmap”,to quantify the cleaning results through quantitative cogitation;(3)Implementation and testing of data visualization and cleaning system for traffic position data.Combining data cleansing and visualization technology,the technical implementation of this article will be practical and effective from requirements analysis,functional module design,and specific implementation technologies,and use the massive taxi data of Hangzhou as test data to analyze the validity and accuracy of the system.This article combines data cleaning and visualization technologies to effectively process and visually analyze traffic data,and use self-designed maps to evaluate cleaning results,making the visual results of traffic conditions accurate and effective.It is hoped that in the future research,the cleaning process can be visualized,so that the source of the specific problem data and the processing method can be seen.The batch data can be centrally processed to improve the data cleaning efficiency and the accuracy of the results.