Data Visualization and Exploration

ggplot II

Review

Last time we introduced the ggplot package (philosophy).

Special assignments¹ (1/2)

Identify a “research” topic in an area of personal interest.

1. Research common visualization types or styles in that area.

• Find one (or more) examples of interest.
• As you did recently, create a document that contains your example visualization(s).
• Attempt an honest discussion the effectiveness of the visualization(s).
• As appropriate (do one or both of),
  • suggest improvements to the visualization based on your class experiences.
  • identify elements you would like to incorporate in your own work.

1. Informal introduction.

Special assignments (2/2)

Considering some research topic in an area of personal interest.

2. Locate a source or sources of publicly-available data¹.

• Searching on the internet (e.g., https://github.com/datasets, https://datadryad.org, https://kaggle.com²) find a total of three datasets with some overlap to your interest.
• Find available documentation to learn about the variables and the history of the data.
• Soon this will be developed into your “semester project” (i.e., “final project”), but begin searching now.

1. If you are a current researcher, you can use your own data.

2. Do not use provided code, only available data.

Goal for Today

We will revisit with our “graph groups” from a few weeks ago.

• Reflect on, assess, and edit (as necesary) the graphs you produced together.
• Ensure that all group members are comfortable with the context, content, and code of the graph.
• Regenerate your graph using ggplot() tools, if possible.
• Other – consider incorporating new datasets or visiting with other groups.

These will be shared by the end of the week to D2L and used as opening slides next week for a brief, relatively informal, Data Viz “Fashion Show”.

Brief content reminder, a break to work, then “normal class”.

“plot annotation”

The idea of annotation as a layer was another fundamental feature of the “grammar”.

labs() (axes and titles)

One more helpful thing to consider at this point would be to use

... + labs(x = ..., y = ..., 
           title = ..., subtitle = ...)

as the analog to mtext() from base R.

With this you can do meaningful annotation.

geom_text() (in-plot text)

After that you might need to annotate (using text) within the space of the graph.

... + geom_text(mapping = aes(label = ...), ...)

• The RHS of label = ... should be a variable name.
• The optional arguments after could include, among others,
  • hjust = ... with 0 (left-justified) or 1 (right-justfied) as options.
  • nudge_x = ... (or _y) as numbers for text placement adjustment.
  • parse = ... with TRUE to interpret text as (porentially)hopefully) properly-formatted mathematical expressions.

Used in combination with geom_text(), geom_label() may make text more readable by providing a “background”.

... + geom_label(mapping = aes(label = ...), ...)

annotate() (open-ended)

Alternatively, annotate() with a variety of options to do many things¹.

... + annotate("geom", x = ..., y = ..., ...)

Here "geom" could be replaced (literaly) by one of,

• "text", with x, y, and label
• "rect", with “corners” fully-specified by xmin = ..., xmax = ..., ymin = ..., ymax = ...
• "segments", with “endpoints” fully-specified by x = ..., xend = , y = ..., yend = ...
• "point", with x, y, and likely color and size

1. Many of these things were provided (in base R) by polygon(), segments(), text() or some combination, just sayin’.

Other ideas

Including just the list of "geom" choices that are compatible with annotate() (e.g., "curve" for curved annotation arrows), there are a lot of special functions for individual tasks.

Feel free to start browsing online resources like https://ggplot2-book.org/ (if you haven’t been) for examples.