Visual Data Mining

In contrast to standard statistical analysis programs, visual mining allows users to interactively explore data using helpful graphs, charts, or interfaces that may employ colors, filters, animation, and different levels of detail. IVEE's Spotfire 1.0, for example, can represent data in a 2-D scatter plot. Along w ith this representation, it creates a number of sliders, one for each statistical variable; these sliders can be used to manipulate the data representation.

The Spotfire screen shows a customer database of a department store. Each colored marker represents a customer, colored by gender, with marker size conveying the amount of money spent on clothing. You can select details of particular customers by clicking on the marker. To filter only those customers who have purchased clothing for at least $500, for example, you simply move the corresponding slider to the right. The second screen displays this selection. Simultaneously, the screen indicates that most of these customers are older than 45 (x axis), two-thirds of whom are women (red), and most have spent more than $2500 (y axis). It's that easy to identify a clear market segment in a customer database.

---

Department Store Database

screen_link (40 Kbytes)

---

Filter Your Se lection

screen_link (33 Kbytes)