交通行動服務MeNGo定期票之電子票證大數據分析與顧客分群細緻化研究

A Study on Refining Customer Segmentation for MeNGo Monthly Passes of MaaS by Using Electronic Ticket Big Data Analytics

自2016年起，高雄市開始推動都會型MaaS (mobility as a service)示範計畫，提供多種付費組合方案。在介接交通部運輸研究所提供之MeNGo定期票會員電子票證資料後，本研究先確認原始資料欄位及可用性，再建立一個大數據清洗、檢驗及串聯流程，並探討MeNGo定期票顧客使用行為之差異。最後，參考國內外電子票證顧客分群之相關文獻，採用使用金額、運輸行為、空間距離等三構面與14個分群特徵變數，結合主成分分析與K-means之分群法進行顧客分群，並以統計檢定進行驗證，再依各分群特徵予以命名。經分群結果得知，MeNGo顧客在定期票使用上確有差異。因此，本研究針對不同顧客分群提出「運具拆售、分區計價」之改良建議方案，以滿足不同顧客需求，應可有效增加MeNGo定期票之銷售量。

A type of Mobility as a Service (MaaS) project in a metropolitan area entitled Mobility as a service policy -MeNGo has been promoted in KaoHsiung City to provide multiple modes of monthly passes since 2016. Having interfaced with MeNGo membership database provided by Institute of Transportation of MOTC, this study firstly verified data columns and availability of raw data from iPASS cards to conduct electronic ticket big data analytics. This study then established a standard process including data cleansing, data verification and data synchronization to explore MeNGo monthly pass customer characteristics and to identify differences among them. MeNGo monthly pass usage patterns which can be summarized as a total of 14 characteristic variables in three dimensions (ticket expense, travel behavior, travel distance) are used to refine customer segmentation with a two-stage clustering analysis combining PCA and K-means method. Statistical test results are also verified to name customer segments correspondingly. With these significant customer segments, product improvement alternatives for increasing sales revenue of MeNGo monthly pass including flexible fares by mode combinations and flexible fares by zones are proposed to satisfy various customer needs.