Crossroads - Spring 2009

Amit Prakash Sawant, Christopher G. Healey, Dongfeng Chen, and Rada Chirkova

Figure 4: Scatter Plot Visualization One: month of purchase mapped to the x-axis, year of purchase mapped to the y-axis, total price mapped to hue, size, and height, number of item types mapped to orientation, and total quantity mapped to luminance and transparency. Here, order ID person173 is highlighted.

In the visualization in Figure 5, we can conclude the following:

1. Few purchase orders were placed in the months of January, July, and October.

2. Many purchase orders were placed in the months of February and December. Also from the visualization, it is evident that the purchase orders in February were predominantly placed around Valentine’s Day (February 14). For the month of December, the purchase orders were placed in the second half of the month, due to Christmas and New Year’s Day.

3. A moderate number of purchase orders were placed in the months of May and June, due to Mother’s Day, Memorial Day, and Father’s Day.

4. There is no particular trend for the shipping type. It is spread out between ground, express, and priority.

Spiral Visualization

In the display in Figure 6, we can make the following observations:

1. In the spiral visualization, the number of turns of the spiral represent a year (we are representing six years: 2002 to 2007). The circular plane is divided into 12 equal sectors (slices) representing different months in a year. Each sector contains six arcs on which objects representing purchase orders are placed.

2. Similar information, as in Figures 4 and 5, can be observed from the above spiral visualization.

Figure 5: Scatter Plot Visualization Two: month of purchase mapped to the x-axis, day of purchase mapped to the y-axis, total price mapped to hue, size, and height, number of item types mapped to orientation, and total quantity mapped to luminance and transparency.

3. Years 2002 and 2007 had far fewer purchase orders compared to years 2003, 2004, 2005, and 2006.

4. Few purchase orders were made in the months of January, July, and October. Many purchase orders were made in the months of February and December.

User Interactions Our system allows the user to interact with the visualizations by translating, rotating, and zooming the environment. Users can change data attribute-visual feature mappings using click-and-drag sliders. Users can also select individual data elements to display a pop-up balloon that describes the exact attribute values encoded by the element.

Conclusions

We have used perceptual visualizations to display Amazon.com purchase orders data. We have mapped visual features, such as spatial position, color, and texture properties to individual attributes of each purchase order. The resulting visual interface allows viewers to accurately explore and discover within the abstract Amazon data. Our method has the ability to adapt to other data sets, for example, a movie database [ 9]. We plan to conduct simple experiments to validate our design choices.