Line Charts

Idiom Line Charts
What: Data One quantitative value, one ordered attribute
How: Encode Points with connection marks between them
Why Show trend
Scale Hundreds of levels of the ordered attribute

Line charts displays one value attribute and one key attribute in a  2-D space, while showing categorical attribute using colors or shapes. It uses one axis for a quantitative attribute and the other for an ordered (sequential/divergent) attribute.

An important feature of line charts is that they also use connection marks to emphasize the ordering of the items along the key axis by explicitly showing the relationship between one item and the next. Thus, they have a stronger implication of trend relationships, as well as continuity.

Line charts should be used for ordered values but not categorical values. A line chart used for categorical data violates the expressiveness principle, since it visually implies a trend where one cannot exist. This implication is so strong that it can override common knowledge.

The underline message of the charts ought to address the difference in heights versus genders. However, the line chart for the above categorical data provides the misleading information such as “The more male a person is, the taller he/she is”.

When designing a line chart, an important question to consider is its aspect ratio: the ratio of width to height of the entire plot. While many standard charting packages simply use a square or some other fixed size, in many cases this default choice hides dataset structure. The relevant perceptual principle is that our ability to judge angles is more accurate at exact diagonals than at arbitrary directions.

The classic sunspot example dataset! The blue line graphs the data itself, while the red line is the derived locally weighted regression line showing the trend. The aspect ratio close to 4 in (a) shows the classic low-frequency oscillations in the maximum values of each sunspot cycle. The aspect ratio close to 22 in (b) shows that many cycles have a steep onset followed by a more gradual decay. 
* Sources: Visualization Analysis and Design, Tamara Munzner, 2014
Show Comments