運用顧客分眾模型及機器學習分析大眾運輸轉乘之需求型態及旅客搭乘特性

近年以來隨著溫度上升，全球暖化（Global Warming）之議題逐漸受到重視，綠色運輸、低碳運輸的概念興起，根據交通部計處2017年「民眾日常使用運具狀況調查」結果顯示，全台灣使用公共運輸之市占率為18.2%，而使用私人運具為70.6%，因此瞭解大眾運輸之需求是項重要的研究議題。本研究針對臺北都會區之臺北捷運與YouBike以2017年1月至2019年7月間的悠遊卡旅次交易紀錄進行分析，研究首先運用大數據分析－統計描述的方法，針對臺北捷運搭乘與YouBike騎乘各類使用者進行空間及時間特性分析，再以悠遊卡之卡號進行判斷YouBike旅次是否轉乘捷運，並探討轉乘行為特性、轉乘旅次鏈與轉乘使用者之需求，最後以RFM模型以K－Means群集分析將屬性相似之使用者進行群集。藉此瞭解臺北捷運及YouBike之熱門站點、路線以及相關YouBike租借使用時間，且更進一步探討由YouBike轉乘臺北捷運之顧客客群，瞭解客群與客群之間的差異。研究結果顯示：（1）臺北捷運之熱門站點與熱門路線皆為於板南線上；（2）YouBike之熱門租賃站與熱門路線皆位於捷運站、商圈與學校周邊；（3）YouBike轉乘捷運受天氣之影響，氣溫增高降雨量越多皆會導致使用者之意願；（4）建構RFM模型與K－Means群集分析，找出顧客群之價值與貢獻度。透過瞭解臺北捷運、YouBike及YouBike轉乘捷運，將可以提供未來大眾運輸系統運輸需求型態之措施與策略之參考。

關鍵字

大數據；悠遊卡；公共自行車；使用者行為；轉乘行為；旅次鏈； RFM模型；群集分析

並列摘要

In recent years, as the temperature has risen, the topic of global warming has gradually received attention, and the concept of green transportation and low-carbon transportation has emerged, according to the results of the 2017 "Survey of People's Daily Use of Vehicles" by the Ministry of Transport. The market share of public transportation in Taiwan is 18.2%, while private transportation is 70.6%. Therefore, there has been strong interest among academics and practitioners in understanding the needs of public transportation. This study analyzes the travel transaction records of the EasyCard between January 2017 and July 2019 on the Taipei MRT and YouBike in the Taipei metropolitan. The primary objective of this paper is to explore the transit behaviours of the Taipei MRT and YouBike rides. In this study, we establish a framework to clean and concatenate datasets in which we examine the space and time characteristics of various cardholders via relevant YouBike rental times, stations and routes of Taipei MRT and YouBike. Then, we apply big data analytics to identify transit patterns, i.e. potential trip chains. Furthermore, we use the RFM model to cluster the users with similar attributes using K-Means cluster analysis and compare results among different segments. The results have shown that (1) the popular stations and popular routes of Taipei MRT are on the Bannan Line. (2) the popular rental stations and popular routes of YouBike are located in MRT stations, business districts and schools. (3) The YouBike transfer to MRT is affected by the weather. The higher the temperature and the more rainfall, the user's willingness. (4) Construct the RFM model and K-Means cluster analysis to find out the value and contribution of the customer group. By understanding the Taipei MRT, YouBike and YouBike Interchange MRT, this study provides a reference for measures and strategies for the transportation demand patterns of the mass transit system in the future.

並列關鍵字

Big Data ； Easycard ； Public Bike System ； Rider Behavior ； Transit Behavior ； Trip- Chain ； RFM Model ； Cluster Analysis

參考文獻

中文參考文獻

Google Scholar

RFM模型與市場區隔，行銷資料科學(2020年4月13日)，檢自：https://medium.com/marketingdatascience/rfm%E6%A8%A1%E5%9E%8B%E8%88%87%E5%B8%82%E5%A0%B4%E5%8D%80%E9%9A%94-74dff44a839c(2021年5月5日)。