Bus Travel Characteristics Of The Elderly Based On Data Mining

As the world’s population structure gradually changes to an ageing population,the proportion of the elderly in public transport travel is increasing year by year.It is necessary for the government to fully understand the difference between the travel characteristics of the elderly and adults,and to provide better public transportation services for the elderly.Accordingly,based on Beijing smart card data,POI data and weather data,this paper studies the travel characteristics of the elderly and adults and the difference in the temporal and spatial distribution of travel passenger flow,and analyzes the travel distance distribution of the elderly with different travel patterns and its main influencing factors.Provide basic support for improving the travel quality of the elderly.The main contents of the paper include:(1)Data preprocessing and multi-modal filling.Use python programming to clean missing data,invalid data,abnormal data,and inconsistent data in the original smart card data.Based on the bus GPS data and the station basic table data,through the data matching method,the bus GPS time-space arrival table is constructed to realize the filling of the spatial information of the boarding and swipe card stations.Through DBSCAN clustering and K-means clustering methods,the card swiping time filling in the two cases of "someone gets off" and "no one gets off" is realized,respectively.And based on the populated data,a conventional bus travel chain is constructed.(2)Analysis of the characteristics of public transport travel among the elderly.Using basic statistical analysis methods,the differences in travel between the elderly and adults on weekdays and weekends were analyzed from the frequency of travel,travel distance,and travel volume.The results show that on weekdays and weekends,the average travel distance of the elderly is 0.7 km and 0.71 km lower than that of adults,respectively,and the average daily bus travel frequency of the elderly is 0.08 times and0.04 times higher than that of adults,respectively.The travel peak of the elderly in the morning coincides with the morning rush hour in the city,and in the evening,they will deliberately avoid the evening rush hour.(3)Analysis of spatial-temporal differences in bus passenger flow among the elderly.First,cluster similar bus stations through hierarchical clustering,and find the centroid of each type of station to generate a Voronoi diagram.Then use the Voronoi diagram to build a directed weighted network.The distribution and spatial heterogeneity of network passenger flow were analyzed by maximum likelihood estimation and the K-S test.And visualize the characteristics of the spatial distribution of the elderly passenger flow.The results show that: the elderly travel more in areas with a higher proportion of the elderly population.Compared with weekends,more older adults will travel to the core areas of the old city and other urban areas on weekdays.The elderly will be more inclined to travel On weekends,they go to suburban parks to play,and the elderly will intentionally reduce their trips to parks and attractions in the old city during weekdays.(4)The distribution of the travel distance of the elderly and its influencing factors.Based on the multiple linear regression model,the travel distance of the elderly with different travel laws is taken as the research object,and the data of time,weather and POI information points are integrated to study the distribution of travel distance and the influence of time,weather and space factors on the travel distance.The results show that the average bus travel distance of the elderly during the peak hours of working days will be longer.Under rainfall conditions,the proportion of older adults who choose public transport for long-distance travel will increase.When the endpoints of bus trips for the elderly are parks,restaurants,clinics,hospitals,and other places,the travel distance will increase significantly.When an endpoint is a place such as a residence or a shopping mall,the travel distance will be significantly reduced.