A1-JonathanChung
From CS294-10 Visualization Fa07
Contents |
[edit] Good Visualization
Source: Business Essentials
[edit] Explanation
This visualization comes from the textbook used in Berkeley's introductory business class (UGBA10), and it appears in the chapter about operations and production.
It is designed to show an example of a supply chain and how a basic commodity such as bread can involve many middlemen. In addition to showing steps, it also mentions the state of the product after each step, so the reader can further imagine what the product would look like at any point in the process.
[edit] Deconstruction
The data model behind this is a relational data model in which records are represented by the names of individual steps in the process (N) and the end result of that process (N). Moreover, the records are ordered in such a way that earlier records appear earlier in the chain (left) whereas later records appear at the end of the chain (right).
The image model in this visualization encodes this data by showing each step in a section and using position to maintain ordering. Arrows are used to connect or map one step to the following step, and above each arrow is the state of the product after that given step.
[edit] Critique
Overall, this is an effective visualization that makes the most of the limited data that it is trying to convey. Simplicity is king a lot of the times, and this visualization, with its clear graphics and layout, does the job well.
[edit] Bad Visualization
Source: Software Engineering (Pressman)
[edit] Explanation
This table comes from one of the textbooks used in Berkeley's Software Engineering course (CS169) and appears in the chapter on project management.
It is designed to summarize and aggregate various software defects in a project in a way that a project manager could easily read. Unfortunately, this table is all but easy to read due to many design flaws.
[edit] Deconstruction
Data is encoded and mapper to a basic table. This table is actually a combination of 4 tables, one each for Serious/Moderate/Minor defects and one for total defects. Spatially, everything within one given row corresponds to one type of defect, and all serious/moderate/minor defects are grouped into their own columns (or more accurately, sub-tables).
Each record represents one type of defect (N), and it keeps track of the number of occurrences of that defect (integers) and the percent of defects that were of that type. Finally, the table provides some aggregation of the data at the bottom in the form of total numbers and total percentages (redundant given that they should all add up to 100%).
[edit] Critique
This is a flawed visualization that is poor on many levels. First, it is overloaded with data, making it difficult to interpret. It's a shining example of a case in which many tables were forcibly combined into one mega table.
Another major problem for one who wants to skim the table to grab a "bottom line" is that if one reads the chart from left to right, row by row, it's disconcerting to see that the numbers add up properly, but the percentages don't (add up to 100). This is because the table actually consists of 3 tables put side to side, but the 3 tables numerically, have nothing to do with each other.
The cryptic abbreviations are unexplained without a legend to refer to.
[edit] Reconstruction
A visualization like this should be broken up to make the data easier to grasp. To do this, I opted to represent the data in a set of 4 pie charts, one for each of the 3 sub categories and one final one that combines the 3. This makes the relative percentages much easier to grasp.
