消費者交易資料之視覺化分析系統

A Visual Analytic System for Consumer Transaction Data

黃品瑄(Ping-Hsuan Huang)

指導教授 ： 林文杰

國立交通大學/資訊學院/資訊科學與工程研究所/碩士(2018年)

摘要

在各式各樣的商業平台當中，消費者之交易資料分析已是一門顯學
——消費模式、顧客行為與特徵已成為進行行銷決策時的重要參考。隨著資料的大量增加與客製化行銷手段的出現，如何有力、且有效率地在消費紀錄中找出更多隱藏資訊、以及找出適當的目標客戶已成了現金財經分析專員的一大困境。再者，基於傳統資料管理系統的分析技術已經漸漸無法跟上迅速日新月異的行銷需求。因此，我們決定呈現一個資訊視覺化系統使資料探索過程更容易，以更進一步提升客戶關係管理之品質。我們提出的系統採用連動式多重圖表視覺化界面，並搭配上視覺化設計與互動，以便於分析人員分群客戶、觀察感興趣的子集合、和嘗試理解資料集或資料向度之間的關聯性。更甚者，我們在系統中加入了分佈視圖，一個將客戶進行自動化降維後之分群結果進行視覺化展示的視圖，與其附上之輔助性的互動功能，它可以為分析多維度資料的關聯性提供另一個更迅速且清晰的視角。為輔助分析人員發揮他們的專業知識，在連動式多重圖表之視覺化界面的視圖亦可自由的被分析人員調整，這使我們的系統提供更高的自由度。我們以台灣知名之百貨公司集團的消費者交易資料作為基礎進行使用範例分析和案例分享，並以此驗證該系統之可用性。

關鍵字

互動式視覺分析 ； 消費資料分析 ； 連動式多重圖表 ； 關聯性視覺化 ； 資料降維

並列摘要

Consumer transaction data analysis has become an important issue within commercial platforms. Detailed shopping patterns and customers’ behaviors and characteristics have already been used to make marketing strategies. Given the increasing amount of data and the rising complexity of marketing promotions, data analysts are now facing the challenges of how to find hidden trends and identify target customers correctly, effectively and efficiently. Traditional analysis techniques based on data management platforms are difficult to catch up with the ever changing marketing requirements. In this thesis, we present an information visualization system to facilitate the data exploration process, and further improve the performance of customer relationship management. Our coordinated-multiple-view system helps analysts to find customers with similar behaviors, observe characteristics of interesting subsets, and figure out the correlation between data attributes. In addition, a distribution view is embedded in our system, which visually demonstrates automatic consumer clustering results generated by dimensional reduction algorithm, with several supportive interactions to further offer a quicker and clearer perspective of relationship among consumers’ multivariate data. We validate our work with the consumer transaction data from a famous department store chain in Taiwan. A set of use cases and findings are provided to show the usability of the system.

並列關鍵字

Interactive Visual Analysis ； Transaction Data Analysis ； Coordinated Multiple View ； Correlation Visualization ； Dimensional Reduction

參考文獻

[1] Tala Mirzaei and L Lyer. “Application of predictive analytics in customer relationship management: a literature review and classifi cation”. In: Proceedings of the Southern Association for Information Systems Conference. 2014, pp. 1–7.

[2] Michael J Shaw et al. “Knowledge management and data mining for marketing”. In: Decision support systems 31.1 (2001), pp. 127–137.

[3] Mollie Cameron and Leigh Nichols. “Understand “Customer Tempo”to Improve Marketing Results”. In: A fi rst data white paper (2008).

[4] Cong Xie et al. “VAET: A visual analytics approach for e-transactions time-series”. In: IEEE Transactions on Visualization and Computer Graphics 20.12 (2014), pp. 1743–1752.

[5] Raghda M Nasr, Sahar Nagaty, and Mohammed B Senousy. “Survey and challenges of mining customer data sets to enhance customer relationship management”. In: IJCA (0975-8887) Vol-107, No-13, PP-34-40 (2014).