成都市快速公交乘客刷卡数据研究

发布时间：2018-05-11 20:51

本文选题：快速公交 + 数据分析　；参考：《西南交通大学》2017年硕士论文

【摘要】：快速公交系统大多采用闸机刷卡,进站乘车的乘车方式,而且车辆灵活运营,较之普通公交,车辆的运营信息通常是不易获取的信息。本文在仅依靠快速公交乘客刷卡数据的情况下,对车辆运营信息进行推算,并根据推算所得数据进行统计分析,对公交服务效果作出评价,并对公交运营管理提供相关建议。本文使用Matlab 2012b进行数据处理和算法构建和执行,使用SPSS Statistics 22进行站点客流统计分析。本文创新性研究工作可概括为以下三个方面:第一,在缺乏GPS等其他类型数据的情况下,仅基于成都市快速公交一周内的乘客刷卡数据,在进行基本数据整理和统计分析过后,通过聚类算法推算出车辆运营信息,包括各个时间段运营的车次数量、车辆运营时刻表以及各个车次在沿线站点的上下车人数等难以获取的数据信息。第二,根据车辆运营信息计算并统计各项满载率指标,然后基于车辆满载情况进行客流流向和客流量统计。并且通过集散量的统计分析和系统聚类方法选出各区域代表站点,提出以代表站点预测其他站点集散量的客流预测方法。并且选用其他日期的数据与预测数据进行了对比,证明了结果的可行性与准确性。第三,根据车辆运营信息计算乘客等车时间、乘客乘车过程中经历的平均满载率和乘客平均每公里通勤时间等对公交乘客满意度有重要的几个指标。对各个指标进行统计分析过后,用K均值的方法将乘车感受不同的乘客分为若干类别,识别出乘客满意度较差的乘客。与传统调查问卷的方式不同,直接通过数据量化指标可更利于公交优化研究,对于公交系统运营更具管理意义。
[Abstract]:The bus rapid transit system mostly uses the gate machine to swipe the card, enters the station to take the bus the way, moreover the vehicle is nimble operation, compared with the ordinary public transport, the vehicle operation information is usually difficult to obtain the information. Based on the data of bus Rapid Transit (BRT) passengers swiping cards, this paper calculates the operation information of the vehicles, and makes statistical analysis according to the calculated data to evaluate the effect of bus service, and provides relevant suggestions for bus operation management. In this paper, Matlab 2012b is used for data processing and algorithm construction and execution, and SPSS Statistics 22 is used for statistical analysis of station passenger flow. The innovative research work in this paper can be summarized as follows: first, in the absence of GPS and other types of data, only based on the data of passengers swiping cards within one week of Chengdu bus Rapid Transit, after the basic data collation and statistical analysis, The operation information of vehicles is calculated by clustering algorithm, including the number of trains in operation in each time period, the running schedule of vehicles and the number of people on and off each train at the station along the route, and so on, which are difficult to obtain. Secondly, according to the vehicle operation information, the full load rate index is calculated and counted, and then the passenger flow direction and passenger flow statistics are carried out based on the vehicle full load situation. Through the statistical analysis of the distribution amount and the systematic clustering method, the representative stations of each region are selected, and the passenger flow forecasting method is put forward to predict the distribution of other stations on behalf of the stations. The feasibility and accuracy of the results are proved by comparing the data of other dates with the predicted data. Thirdly, according to the vehicle operation information, the passenger waiting time, the average full load rate and the average commuting time per kilometer have several important indicators for bus passenger satisfaction. After the statistical analysis of each index, the passengers with different sense of travel are divided into several categories by means of K-means method, and the passengers with poor passenger satisfaction are identified. Different from the traditional questionnaire, it can be more beneficial to the study of bus optimization by directly quantifying the data, and it has more management significance for the operation of the public transport system.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F572.88

【参考文献】