Research On Traffic Trend Visualization Method Based On Trajectory Big Data

With the rise of smart cities,the number of taxis in cities continues to increase,resulting in massive taxi track data.Through the study of these massive trajectory data,it is helpful for professionals to explore some rules and obtain more accurate and real analysis results,so as to propose improvements to urban road planning and provide some powerful decision-making support for taxi drivers and location service providers.Big data visualization technology can display massive data in a hierarchical form of interaction,and can analyze trajectory data more intuitively and concisely and find the laws and characteristics in the data.Therefore,this paper takes the trajectory data of "Didi Chuxing" in Xi’an as the research object,and uses the clustering algorithm to mine the massive trajectory data,and on this basis,the method of data visualization and analysis is adopted,and the travel characteristics of taxis and their temporal and spatial changes are excavated through the visualization method of multi-view collaboration and map combination.The main research content of this paper includes the following aspects:(1)The original trajectory data is preprocessed,the outlier filtering algorithm is used to remove the out-of-bounds data and abnormal data in the GPS trajectory data,and the coordinate system conversion algorithm is used to convert the trajectory data set into coordinates.(2)An improved clustering algorithm LAND-DBSCAN is designed to view the shortcomings of DBSCAN clustering algorithm on parameter sensitivity,data mining of "Didi Chuxing" trajectory data in Xi’an,excavating the main congestion areas in roads in different time periods and visual analysis through heat maps,and verifying the accuracy of clustering through multiple experiments.(3)An improved clustering algorithm AKD-means is designed for the shortcomings that the number of clusters in K-means algorithm requires manual input and is easily affected by noise points,and the hot spots of taxi passengers are excavated and visually analyzed.The experimental results show that the clustering algorithm designed in this paper can avoid the uncertainty of experimental results caused by manual input parameters,and the excavated hot spots have higher accuracy.(4)Design and implement a visualization system based on trajectory data analysis.The trajectory dataset of "Didi Chuxing" in Xi’an was visually analyzed from the taxi trajectory module,hot spot module,vehicle origin point module and heat map module.