Visualization and Analysis of Clickstream Data of Online Stores for Understanding Web Merchandising

  • Authors:
  • Juhnyoung Lee;Mark Podlaseck;Edith Schonberg;Robert Hoch

  • Affiliations:
  • IBM T.J. Watson Research Center, P.O. Box 218, Yorktown Heights, NY 10598, USA. jyl@us.ibm.com;IBM T.J. Watson Research Center, P.O. Box 218, Yorktown Heights, NY 10598, USA. podlasec@us.ibm.com;IBM T.J. Watson Research Center, P.O. Box 218, Yorktown Heights, NY 10598, USA. ediths@us.ibm.com;IBM T.J. Watson Research Center, P.O. Box 218, Yorktown Heights, NY 10598, USA. rhoch@us.ibm.com

  • Venue:
  • Data Mining and Knowledge Discovery
  • Year:
  • 2001

Quantified Score

Hi-index 0.01

Visualization

Abstract

Clickstreams are visitors' paths through a Web site. Analysis of clickstreams shows how a Web site is navigated and used by its visitors. Clickstream data of online stores contains information useful for understanding the effectiveness of marketing and merchandising efforts, such as how customers find the store, what products they see, and what products they purchase. In this paper, we present an interactive visualization system that provides users with greater abilities to interpret and explore clickstream data of online stores. This system visualizes the effectiveness of Web merchandising from two different points of view by using two different visualization techniques: visualization of sessions by using parallel coordinates and visualization of product performance by using starfield graphs. Furthermore, this system provides facilities for zooming, filtering, color-coding, dynamic querying and data sampling. It also provides summary information along with visualizations, and by maintaining a connection between visualizations and the source database, it dynamically updates the summary information. To demonstrate how the presented visualization system provides capabilities for examining online store clickstreams, we present a series of parallel coordinates and starfield visualizations that display clickstream data from an operating online retail store. A framework for understanding Web merchandising is briefly explained. A set of metrics referred to as imicro-conversion rates, which are defined for Web merchandising analysis in our previous work (Lee et al., Electronic Markets, 2000), is also explained and used for the visualizations of online store effectiveness.