EOW04: Survey Answers

Examples of correct answers to this week’s survey

Example Answers to The Questions

Explain the encoding of a bubble chart.

A bubble chart encodes a value to area/size, this specific bubble chart encodes median houshold income for each state as area/size

Why might the bubble chart be a bad choice? (you can focus on encoding principles)

Area is not a great encoding when we want to compare values or retrieve a specific value (it is really hard to determine what the median income of a state actually is based on size and is hard to compare the median income across states). In addition, look up is difficult because we don’t appear to have an intentional/purposeful positional encoding that would help us find a specific state.

A bubble chart is probably not a good choice for this particular data set - but there are times when it might be appropriate. Can you think of cases where those advantages are useful?

*If you wanted to focus on large differences or draw attention to outliers, a bubble chart could be useful because it is relatively easy to tell which bubbles are biggest/smallest and your eye is drawn to them. Another usecase is if your data had more than two dimensions, the area encoding of a bubble chart could be used to encapsulate an additional dimension on top of other (idealy positional) encodings. *

A treemap is also inappropriate for this data set. Why?

Treemaps are made specifically for heirarchial or part-whole data, this data is not heirarchial/part-whole.

Describe a situation where you had data for each state where a treemap would be appropriate.

It needs to be part whole (so the values add up to a whole that we care about). A treemap is most useful if there is some hierarchy that is interesting.So an example: we have the population per state, and include a grouping by region.

Describe the encoding (there are actually many variables) - it probably helps to describe the data (list the variables that are presented)

Color encoding: the ‘Daily/Forecast Temperature’ and ‘Average High/Low’ categories are encoded by color.

Position encoding: The dates are encoded along the x-axis and the temperatures are encoded along the y-axis, individual temperature values (for the daily high, low, average high, and average low) are encoded through vertical and horizontal position along these axes

Length: the difference between the high and low temperatures each day is encoded by the length of the bar between them (this is somewhat of a redundant encoding)

Tilt: the change in average temperatures from day to day is encoded in tilt or orientation of the line connecting them