Research On Extraction And Analysis Of Taxi Passenger Travel Characteristics Based On Big Data

With the development and wide application of mobile network and communication technology,these technologies can be used to obtain the GPS data during the taxi is running.Analyzing the passenger travel characteristics from these GPS data,the urban traffic and the taxi industry managers can make decisions in urban transportation planning,urban traffic flow equilibrium and vehicle scheduling.However,in the face of growing data,research on how to analyze quickly and deal with these massive taxi GPS data is also a hot topic of current large data research.Big data is widely concerned in the field of research in recent years,a variety of large data processing platform framework was put forward,which the most popular is Hadoop and Spark.Spark has the advantages of high efficiency,low cost,simple data processing.Compared with the Hadoop platform,it has overcome the MapReduce shortcomings of high delay and long running time.Therefore,this paper chooses Spark as a large data parallel computing framework for taxi GPS data processing,which is used to extract and analyze the taxi passenger travel characteristics.The main contents of this paper are as follows: firstly,based on the analysis of the shortcomings of traditional residents' travel characteristic acquisition method,this paper gives the method of extracting taxi passengers' travel characteristics by using taxi GPS track data,and analyzes its advantages and disadvantages.Secondly,this paper introduces the technology of extracting and analyzing the passengers' travel characteristics based on the GPS track data of the taxi.The pretreatment method of GPS data was proposed,a map matching algorithm based on dotted line was implemented.Then based on the analysis of Spark principle and its operation mechanism,a number of extraction algorithm of passenger travel characteristics based on Spark platform were proposed,and the performance of these algorithm were discussed.Finally,based on the real history of the taxi GPS track data,the urban taxi passengers' travel characteristics were extracted and analyzed,include travel time distribution characteristics of passenger,spatial distribution characteristics of taxi and the time and spatial distribution characteristics of empty-loaded rate.Combined with the distribution of urban population and the daily travel law of the public,the travel characteristics were analyzed and explained.Results show that the combination of Spark framework and GPS data mining method can quickly analyze the characteristics of taxi travel characteristics.