Intro to Data Science

Lab 6 – Visualization II

A Guide to Your Process

Scheduling

Learning Objectives

Practice

Supporting Information

Class Discussion

Today’s Plan

• Muddiest Point Review
• Data Visualization with ggplot2
  • Editing theme elements
• Multi-Panel Graphs
• GitHub Presence Check-Ins (1-on-1)
  • Not graded! Don’t stress!

Today’s Learning Objectives

After today’s session you will be able to:

• Modify background elements in a ggplot2 graph
• Create publication-quality figures with ggplot2
• Explain the difference between plot faceting and plot grids

Muddiest Point Review

• Recurring topics from most recent MPs:

• What other topic(s) would you like to review?

ggplot2 Review

• ggplots require: (1) data, (2) aesthetics, (3) geometries
  • Optionally can mess with theme parameters

Theme Internal Structure

• Theme is composed of elements

• Elements can be modified as desired inside of theme function

• Each type of element has a different ‘helper function’ needed to modify that element
  • Change text = use element_text
  • Change line = use element_line
  • Remove an element with element_blank

Theme Syntax

• You use the theme function once with as many element_... functions as you need

• Here’s an example of the proper syntax

# Make a simple scatterplot
ggplot(data = my_df, mapping = aes(x = x_var, y = y_var)) +
    geom_point() +
    # Modify its theme to make the axis font size bigger
    theme(axis.text = element_text(size = 20),
          # Also remove the grid lines
          panel.grid = element_blank())

• Note how the theme and element_... functions are used together

Gridline Theme Components

• You’ll learn theme argument names as you work more with ggplot2

• Here are a few broadly relevant ones:
  • Gridlines = panel.grid
  • Plot background = panel.background
  • Axis lines (X & Y) = axis.line

Get Ready

1. Create a script for this week

2. Load ggplot2

3. Read “minnow.csv” into R and check the structure!

4. Copy the final graph we made last time

5. Assign the graph to an object

Remove Gridlines

• To that graph, add the following code:
  • theme(panel.grid = element_blank())

• What does this do to your graph?

• What happens if you add these two lines as well (inside of the theme parentheses!)?
  • panel.background = element_blank()
  • axis.line = element_line(color = "black")

Remove Gridlines

Changing Text Size

• We can also modify text size inside of theme

• Axis “title” vs. axis “text”
  • axis.title = axis label text (given to labs function)
  • axis.text = text on axis tick marks

• Want to modify just X or Y? Add that to the argument name!
  • E.g., theme(axis.text.x = element_text(...))

Change Text Size

• Take your most recent graph
  • No gridlines, no background gray square, black axis lines

• And make the following tweaks:
  • Make the axis title font size 15
  • Make the axis text font size 13

• What does that leave you with?

Change Text Size

Customizing Legend

• You can also customize the plot legend in the theme function!
  • Legend placement = legend.position
  • Legend title = legend.title

• legend.position works differently than most other elements!
  • Instead of wanting element_... it wants c(<x position>, <y position>)
  • Positions range from 0 (left / bottom) to 1 (right / top)

Customizing Legend Syntax

• Check out an example where we put the legend in the middle of the plot

# Example scatterplot
ggplot(data = my_df, mapping = aes(x = x_var, y = y_var)) +
    geom_point() +
    # With legend in the middle of the space
    theme(legend.position = c(0.5, 0.5))

• That graph would have the legend in both:
  • the center (left / right)
  • the middle (top / bottom)

Customize Legend

• To the graph you created last practice:
  • Remove the legend title
  • Experiment with legend placement until you’re happy

• You may put the legend wherever you’d like but:
  • It should not overlap any points / boxplots

• What does that graph look like?

Customize Legend

Temperature Check

How are you Feeling?

Multi-Panel Background

• Sometimes nice to have multiple graphs next to each other
  • Makes direct comparison easier
  • Journals have limits on number of figures but multi-panels still count as 1

• Two methods (for ggplots):
  1. ggplot2::facet_grid()
  2. cowplot::plot_grid()

Facted Graphs

• ggplot2 has an internal way of handling this called facets

• Facets work similarly to geometries
  • You get separate plots for each level of the facet variable

• Facets must all be the same plot type and have identical axes
  • Sometimes not an issue but good to keep in mind!

Fact Syntax

• To facet into 1 row x many columns:

# Example scatterplot
ggplot(data = my_df, mapping = aes(x = x_var, y = y_var)) +
    geom_point() +
    # Facet into rows of some other variable
    facet_grid(. ~ facet_variable)

• To facet into many rows x 1 column:

# Example scatterplot
ggplot(data = my_df, mapping = aes(x = x_var, y = y_var)) +
    geom_point() +
    # Facet into rows of some other variable
    facet_grid(facet_variable ~ .)

Facet Example

ggplot(data = penguins, aes(x = body_mass_g, y = flipper_length_mm, color = species)) + 
  geom_point() +
  facet_grid(. ~ island) +
  labs(x = "Body Mass (g)", y = "Flipper Length (mm)") +
  theme(legend.position = c(0.87, 0.85),
        legend.title = element_blank(),
        panel.background = element_blank())

Facet

• Using the fish data, make a new graph that:
  • Has nest diameter on the X axis
  • Has nest depth on the Y axis
  • Is a scatterplot
  • Faceted by species
  • Plus any additional theme tweaks you want to make!

• What does your plot look like?

Facet

Plot Grids

• Facets work great when all panels are the same
  • What about when you want different graphs in each panel?

• cowplot::plot_grid lets you put multiple different graphs together

• Have to make graphs separately first, then combine them

• Example syntax:

cowplot::plot_grid(plot1, plot2, ncol = 1, nrow = 2)

Plot Grid Example

Plot Grids

• Make two graphs using the fish data:

1. Copy your faceted graph of diameter vs. depth
   • But remove the facet by species
2. Make a boxplot with flow on the y-axis and species on the x-axis

• Using plot_grid, make a multi-panel graph with these two graphs
  • Make them side by side (I.e., 2 columns, 1 row)

Plot Grids

Temperature Check

How are you Feeling?

GitHub Presence FAQ

• Worth 40 pts (16% course grade)

• Checklist-style rubric on Canvas

• Due day before last lab

• Can basically finish all of it now though if you want!

GitHub Presence Assignment

• This assignment will seriously help in interviews / job apps!
  • Demonstrates your data science skills

• I don’t want anyone caught unawares by this assignment
  • So I’ll meet with each of you 1-on-1 today to see where you’re at so far

• Good chance for you to ask any questions you have!
  • Also lets me give you tips for success

Upcoming Due Dates

Due before lab

(By midnight)

• Homework #6
• Muddiest Point #7

Due before lecture

(By midnight)

• Homework #7
• Submit Draft 2 of Function Tutorials
  • Double check rubric to see that you’re not leaving any points on the table!
  • Remember to also submit the Revision Response

GitHub 1-on-1s

• Stick around until we do our 1-on-1

• After you have met with me you can leave
  • Though you’re welcome to stay and work on Homework #6 / course assignments!

• Any volunteers to go first?