Article 1: How to Graph Badly or What NOT to Do
This first article, much like the introduction chapter in the first assignment, detailed what poor graphs are in relation to what should be used in the creation process of graphs. Chartjunk, a term coined by Eduard Tufte, is all the unneccesary add-ons to graphs that serve no purpose in communicating research or data findings. In other words, Chartjunk focuses on many artistic qualities vs. the presentation of clear data. In addition to Chartjunk, there are many inappropriate uses of colors and 3-D graphics that mean nothing within the interpretation of information on graphs. Anothr purpose for this article is to show all the bad qualities of graphs so that anyone who wishes to make successful graphs may do so without error. A system of rules that help identify the negative aspects of graphs are Wainer’s Rules for Bad graphs. These rules clearly lay out very poor ways to distract the readers when trying to read data. Furthermore, a really negative aspect in vizualizing data is called Data-hiding. In comparison to cherry-picking, data-hiding doesnt cherry-pick the data out of the analysis process but the hiding process is when the data is all there like in a correlation, but instead of having the points and the line together representing a possible trend, the points are removed and the line is all that is left. Towards the end of the article I enjoyed learning that the types of shading used in data visualization can impact the way we perceive how one shade impacts the perception of importance from the reader. For example, if you shade one bar in a bar graph darker than all the others, then perceptually, the reader will more likely make an inappropriate inference to conclude the darker bar is more important.
Article 2: The Gospel According to Tufte
The start of this article addressed my interest in the data-to-ink ratio. Essentially, Tufte proposes that one should “above all else show the data, maximize the data-ink ratio, erase non-data-ink, erase redundant data-ink, and revise and edit”. One way to maximize the data-to-ink ratio while emphasizing data in a creative way is to bolden the axis lines and/or highlight the key results in a data set by using stronger colors, not more. This process doesn’t alter the cognitive content of the graph but purely empahszes the data with a slight artistic touch. Grid lines are a way to maximize the data-to-ink ratio without overdoing it. When it comes to showing a 3-D model (appropriately), one can use a surface mesh diagram with a 3-D grid to represent a clear figure to the reader. Something that is quite important regarding this article and the previous is that judgement concerning the data-to-ink ratio is always contingent on the appropriatness of the context of the relationship the data wishes (or the researcher) to represent. Therefore, when you are deciding whether or not to make good data redundent or whether to emphasize certain characteristics in a graph is dependent on the relationship and theme of the data representation. For example, if oen is trying to communicate a trend in a data report, one shouldn’t focus on a possible outlier in the data set. In other words, you wouldn’t want to emphasize an outlier or bolden it in any way that detracts away from the broader theme regarding the motif behind the story of the visualization process of your data. Finally, when someone wants to label cetain areas of a complex graph, like the ear example in figure 2.29, it is best to use the full name of whatever area you are labeling. If parts are not labeled with the full name, a legend actually becomes more complicated with the addition of more parts in a graph that requires a hunt for the correct term in relation to the letter provided in the complex diagram.