Research On Information Mining And Visualization Based On Taxi GPS Data

Urban residents are the main participants of urban traffic,and their travel activities can directly affect the operation of urban traffic.Therefore,the study of urban residents’ travel activities can provide accurate and reliable decision support for urban traffic planning,design,and management.Taxi GPS data contains important spatiotemporal semantic information related to urban traffic.On the one hand,it can reflect the starting point,destination,travel time,travel distance,and other travel information of urban residents,on the other hand,it can reflect the congestion of urban road sections.This paper takes GPS data of taxis in Lanzhou City as the research object.From the perspective of time and space,it mainly studies the peak and trough periods of residents’ travel in Lanzhou City in time and analyzes the evolution of travel hot areas and the potential congestion state of road sections in space.The method of information visualization is designed by using Arc GIS software and reverse geocoding technology.The research mainly includes the following aspects(1)Preprocessing of taxi GPS data: build the database of taxi GPS data,deal with missing data,invalid data,redundant data,etc.,construct the model of taxi boarding event and taxi alighting event,extract the boarding point data and alighting point data(O/D point),design the map matching method based on Arc GIS software,to understand the spatial and temporal distribution of O/D point data,Complete data preparation.(2)Statistical analysis of urban residents’ travel status: using the mathematical-statistical method to analyze the total daily travel volume of residents in Lanzhou,we can have a general understanding of the taxi travel status in Lanzhou.The travel volume and travel distance of each hour are calculated to find out the peak and trough periods and daily travel distance of Lanzhou residents.Finally,the OD matrix of residents in Lanzhou is calculated to analyze the travel status of residents in different urban areas.(3)Urban residents travel hot area mining: a travel hot area mining model is proposed.The model improves the hierarchical clustering algorithm,sets the stop condition of the algorithm,determines the range of the number of clusters according to the relevant knowledge of the traffic district,and further determines the specific number of clusters combined with the contour coefficient.To explore the evolution of travel hot spots in urban areas in one day,the data of six different time characteristics are selected to study,which are 00:00-02:00,08:00-10:00,12:00-14:00,14:00-16:00,18:00-20:00,and 22:00-24:00.The improved clustering model is used to cluster the data of boarding and alighting points(O/D points)in different time characteristics to get the hot areas of residents’ travel in different time characteristics and analyze the evolution of hot areas of residents’ travel in different time characteristics.(4)Congestion section mining: a potential congestion state mining model is proposed,which optimizes the parameter selection method of the DBSCAN algorithm,and uses the optimized DBSCAN algorithm to cluster the morning and evening peak data under the passenger-carrying state,and finds the dense area of taxi aggregation under the passenger-carrying state.This paper constructs the calculation model of the average driving speed of the road section,calculates the average driving speed of the urban road section and the average driving speed of the taxi dense area road section under the condition of carrying passengers.Through comparative analysis,this paper explores the possible traffic congestion in Chengguan District of Lanzhou city.(5)Visualization research: in the stage of hot area visualization,the center point of the hot area is extracted and visualized on Arc GIS.With the help of Alibaba cloud’s inverse geocoding interface,the clustering hot spots are reversely geocoded to analyze the location information of residents’ travel hot spots.In the visualization stage of the congested road section,the location information of the road section is analyzed by using the inverse geocoding method.