透過您的圖書館登入
IP:18.226.177.223
  • 學位論文

悠遊卡大數據應用於大眾運輸乘客旅運型態之研究

A Study on Applying the Big Data of Easy Card to Construct the Passengers' Travel Patterns of Public Transportation

指導教授 : 陶治中

摘要


近年來政府雖積極推動公共運輸相關建設,然而公共運輸市佔率比例始終不如預期。根據交通部統計處資料顯示,我國公共運輸市佔率連續三年雖呈現持續成長狀態,但每年成長幅度仍為逐年遞減,此顯示我國公共運輸服務與民眾實際之旅運需求仍有相當程度的落差,因此大多數民眾仍仰賴私人運具。 有鑑於此,本研究期望以大數據分析來探討使用者之實際旅運行為,並選定目前使用者人數最多之悠遊卡票證資料為大數據資料來源,然後瞭解國內公共運具市場提供服務之特質,再建立通勤族群之旅運特質時空間關聯模式。本研究係依照資料挖掘技術流程,依序將使用者通勤旅次長度及旅次起始時間等旅運特性資料進行分群,以此建立旅運特質之時空間關聯規則,並輔以問卷調查分析結果,進而探究民眾真實之旅運型態,以瞭解民眾對於公共運輸服務認知之缺口。 本研究經由時間分群之結果可得知多數通勤族群接受公共運輸服務之時間分布,同時亦透過通勤距離分群瞭解多數民眾之乘車旅運特性,而檢視出在若干之旅次長度或是起點下,民眾對於運輸系統的選擇偏好。根據研究結果顯示,在旅次長度落於1至5公里的區間內,民眾較偏好選擇公車;在旅次長度高於15公里以上,民眾則較偏好捷運。 本研究進一步進行旅運型態之問卷分析,結果得知在票價優惠、乘車次數上限及里程計費等三個乘車誘因情境,能有效改變機車族群使用公共運具之頻率,而在票價優惠的情境條件能吸引每週騎車通勤1至2次及搭乘大眾運具每週1至2次的族群;在乘車扣款上限的情境下,能吸引所得1萬元以下且幾乎每天騎車通勤的族群;而在里程計費的情境下則能吸引,所得1萬元以下、平均通勤時間為40至50分、每週搭乘大眾運具一至兩次的族群。 綜合上述,本研究透過悠遊卡電子票證乘車交易資料的挖掘,建立通勤族群對於選擇大眾運具之旅運時空間關聯規則,並改善機車族群使用公共運具之頻率。後續研究可再精進探討國內公共運輸服務缺口之分析程序,而提供兼具有效性與可行性之公共運輸服務缺口改善方案。

並列摘要


In recent years central and local governments have deployed public transportation systems; however, the growth rate of public transportation usage is still below the expected level. According to results of Ministry of Transportation and Statistics Department the proportion of public transportation usage has grown in recent three years but the scale of growth proportion decreases annually. It shows that a gap exists between public transportation service supply and travel demand. Therefore, most of people still rely on private transportation instead of public transportation. In order to understand characteristics of public transportation service and to construct relevance models of travel pattern according to time and space for public transportation commuters, this study will focus on analyzing big data of Easycard and clustering commuters by using trip length and trip time. Furthermore, a questionnaires survey is also conducted to explore travel patterns and service gap cognition of commuters in a real case. Results show that most of commuters really travel according to trip time and trip length. It is also shown that commuters will take bus when their trip lengths are between 1 and 5 kilometers, while take MRT when their trip lengths are over 15 kilometers. The results of the questionnaires survey shows three incentive scenarios of concessionary fare, the maximum frequency of taking public transportation and the mileage calculation can effectively change usage frequencies of motorcyclists. Using the scenario of concessionary fare it will attract commuters who drive motorcycle once or twice every week or take public transportations once to twice every week. In the scenario of the maximum frequency of taking public transportation, it will attract commuters who drive motorcycle every week with income per month below 10,000 NTD. While in the scenario of mileage calculation it will attract commuters who take public transportation once or twice every week with income per month below 10,000 NTD and average travel time between 40 to 50 minutes. In summary, this study constructs travel patterns of public transportation commuters by using big data from Easycard database. It is suggested to explore more detailed travel behavior relevance models in future.

並列關鍵字

Big Data Smart Card Data Mining Travel Pattern

參考文獻


4.吳沛樵(2007),非接觸式智慧卡對政府效益之分析,國立臺灣大學土木工程學研究所碩士論文。
6.林良泰、陳乃萁. (2010),K-means集群分析法應用於號誌定時時制時段劃分之研究,運輸學刊第二十二卷第四期。
9.郭奕杖等人(2012),探討高鐵實施不同票種之價格促銷方案對小汽車駕駛人運具選擇行為的影響,運輸學刊第二十四卷第四期。
13.張恭碩(2007),非接觸式智慧卡於停車收費之效益分析,國立臺灣大學土木工程學研究所碩士論文。
14.張智勝(2013),以質性研究方法探究大學生使用機車與大眾運輸工具之態度 -以高雄師範大學為例,國立成功大學都市計劃研究所碩士論文。

延伸閱讀