CPSC 533c Assignment 1

Jeremy Hilliker

"Good" Visualization

Source: "Martha Stewart Better than Google" Economist.com. Sept 1st, 2005.

The chart shows Martha Stewart Living Omnimedia's (MSO) share-price throughout the course of Martha's legal case and notes milestones in the case. Martha is the public face of her company and her image arguably accounts for 90% of the company's valuation. She was publicly investigated for insider trading in 2002, and was sentenced for related charges in 2004.

The graph follows the conventions for depicting share-price relative to time, making it easy for anyone familiar with financials to read. The simple line-graph presentation against linear (and 0-based) axes is also easily understood by laymen.

The milestones in Martha's case which the author chose to note are indicated by a vertical dotted line at their respective times, which are accompanied by a brief but descriptive label. This allows the viewer to both find how these events affected MSO's share-price and to see what may have caused sudden changes in value without cluttering the line-graph with ambiguous symbols or extraneous notation.

"Bad" Visualization

Source: "A New Effort To Analyze Interactive Computing On The Internet" USCD Jacobs School of Engineering. Sept 13, 2005.

This image reputably depicts a social "collaboration network". Each node is an academic author and each edge represents a joint paper between two authors.

This visualization is poor for a number of reasons. None of the nodes are labelled, so the viewer cannot determine who is adjacent to whom. The colours were not well chosen. Both red and green are present, but color blind users may not be able to discern between them. In my opinion, the dark blue colour tends to dominate the view of the graph. Furthermore, the use of colour in the graph appears to have no meaning or pattern.

This graph is only useful for detecting the mass of nodes at its centre. This is presumably a group of highly collaborating authors. These nodes, however, are too tightly packed together to see any detail in that region. It cannot even be determined that they are, in fact, collaborating with each other. Proximity is not enough to decide this since it has little meaning in the outlying areas of the graph.

I think that this visualization is too busy to be effective.