Anne Harding

1. Good Example

We've discussed in class how difficult it can be to convey several different datatypes at once without creating a confusing visualization; I think that this does an excellent job of simultaneously showing multiple factors related to the upcoming Republican primaries. In a political visualization, Americans will immediately associate red with conservative voters and blue with more liberal voters, so the color scale makes sense to represent the voting demographics for the primaries in each state. While it could be difficult to distinguish between some of the middle values, the date bar intersecting the more conservative side from the more moderate side allows us to immediately see which states are more conservative or more moderate. The states are also stacked, with positions further away from the date bar indicating a higher percentage of either conservative or moderate voters, eliminating the need to try to tell which shades of purple are more red or more blue. Size is also used in this visualization to convey the number of delegates associated with each state. Although the actual numbers may be tricky to figure out, comparing the stakes of each state is easy in this format.

Although I think the visualization is great, there are a few things that could be confusing to readers. The color scale is easy to see and makes it easy to compare the relative dominance of conservative vs. moderate voters, but these relative concentrations are not spaced consistently from the horizontal date span bar - the distance from the bar is based solely on the number of other primaries that are occurring on the same day vs. the political makeup of the voter pool, which might take a moment to decipher. The date bar is also not spaced consistently, although I don't think that that's necessarily a problem for the visualization; our concern is not with the measurement of time between the events, but rather with their order, which is shown here uninterrupted by spans of time without primary contests.

2. Bad Example

This visualization shows several different types of impact events, and information about the size of the object impacting earth, the average number of years between impact, the amount of energy expended by an object of each particular size, potential consequences of each type of object impact, and some examples of historical impacts that match these profiles. Neither the axes nor the images themselves give us a consistent, accurate sense of time scale, relative impact size, or amount of energy expended. The images are not at all correlated with the actual amount of energy expended by each type of impact, and because the scales are labeled at inconsistent intervals and often with exponential values, the visualization "lies" in almost every sense. Even ignoring the major "jumps" on each axis, it isn't natural or intuitive for most non-scientific audiences to conceive of time or physical measurements in exponential measures. Additionally, much of the text is difficult to read - particularly the small, bright blue, vertical text, which is hard to understand both due to its orientation and its color.

3. Botanical Tree

I would think that this visualization would be hard for most viewers to interpret. While it's fairly easy to get an immediate sense of relative size to some extent, the fact that the visualization exists in 3-dimensional space makes it hard to evaluate and compare all of the branches simultaneously. Although cones and spheres are easily distinguishable when comparing shapes of the same size, the combination of shapes is less useful with this large variety of sizes. As an example, it was difficult to see that one of the spheres in figure 12 had a number of cones, because the cones were so much smaller than the sphere they were attached to - only in the detail are the small, closely-spaced cones really visible. The color scheme is also a barrier to good visualization, as the large number of distinct colors cause some visual confusion.