MATH167R: Data visualization

Peter Gao

Warm-up

1. What does the following code do?

   library(palmerpenguins)
   library(tidyverse)
   data(penguins)
   penguins |> 
     group_by(species, island) |>
     summarize(mean_body_mass = mean(body_mass_g, na.rm = T))

Warm-up

library(palmerpenguins)
library(tidyverse)
data(penguins)
penguins |> 
  group_by(species, island) |>
  summarize(mean_body_mass = mean(body_mass_g, na.rm = T))

# A tibble: 5 × 3
# Groups:   species [3]
  species   island    mean_body_mass
  <fct>     <fct>              <dbl>
1 Adelie    Biscoe             3710.
2 Adelie    Dream              3688.
3 Adelie    Torgersen          3706.
4 Chinstrap Dream              3733.
5 Gentoo    Biscoe             5076.

Overview of today

• Data visualization in base R
• Data visualization with ggplot2
• Principles of data visualization

Data visualization

Why look at data?

Why look at data?

All of these datasets have the same means for x and y, and (approximately) the same correlation:

# A tibble: 13 × 4
   dataset    mean_x mean_y correlation
   <chr>       <dbl>  <dbl>       <dbl>
 1 away         54.3   47.8        -0.1
 2 bullseye     54.3   47.8        -0.1
 3 circle       54.3   47.8        -0.1
 4 dino         54.3   47.8        -0.1
 5 dots         54.3   47.8        -0.1
 6 h_lines      54.3   47.8        -0.1
 7 high_lines   54.3   47.8        -0.1
 8 slant_down   54.3   47.8        -0.1
 9 slant_up     54.3   47.8        -0.1
10 star         54.3   47.8        -0.1
11 v_lines      54.3   47.8        -0.1
12 wide_lines   54.3   47.8        -0.1
13 x_shape      54.3   47.8        -0.1

Why look at data?

from Jackson (1980) as reproduced in Kieran Healy’s Data Visualization.

What makes a good data visualization?

1. Legibility: Is it easy for the viewer to understand and interpret your visual?
2. Integrity: Does the visual accurately communicate some insight about the data?
3. Beauty: Is the visual appealing and uncluttered?

Other Readings:

• Information is Beautiful

Legibility

from Edward H. Adelson as reproduced in Kieran Healy’s Data Visualization.

Legibility

Legibility

from Kieran Healy’s Data Visualization.

Integrity

Integrity

Example: Movies data

movies <- readr::read_csv("https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2018/2018-10-23/movie_profit.csv")

...1	release_date	movie	production_budget	domestic_gross	worldwide_gross	distributor	mpaa_rating	genre
1	6/22/2007	Evan Almighty	1.75e+08	100289690	174131329	Universal	PG	Comedy
2	7/28/1995	Waterworld	1.75e+08	88246220	264246220	Universal	PG-13	Action
3	5/12/2017	King Arthur: Legend of the Sword	1.75e+08	39175066	139950708	Warner Bros.	PG-13	Adventure
4	12/25/2013	47 Ronin	1.75e+08	38362475	151716815	Universal	PG-13	Action
5	6/22/2018	Jurassic World: Fallen Kingdom	1.70e+08	416769345	1304866322	Universal	PG-13	Action
6	8/1/2014	Guardians of the Galaxy	1.70e+08	333172112	771051335	Walt Disney	PG-13	Action
7	5/7/2010	Iron Man 2	1.70e+08	312433331	621156389	Paramount Pictures	PG-13	Action
8	4/4/2014	Captain America: The Winter Soldier	1.70e+08	259746958	714401889	Walt Disney	PG-13	Action
9	7/11/2014	Dawn of the Planet of the Apes	1.70e+08	208545589	710644566	20th Century Fox	PG-13	Adventure
10	11/10/2004	The Polar Express	1.70e+08	186493587	310634169	Warner Bros.	G	Adventure
11	6/1/2012	Snow White and the Huntsman	1.70e+08	155136755	401021746	Universal	PG-13	Adventure
12	7/1/2003	Terminator 3: Rise of the Machines	1.70e+08	150358296	433058296	Warner Bros.	R	Action
13	5/7/2004	Van Helsing	1.70e+08	120150546	300150546	Universal	PG-13	Action
14	5/22/2015	Tomorrowland	1.70e+08	93436322	207283457	Walt Disney	PG	Adventure
15	5/27/2016	Alice Through the Looking Glass	1.70e+08	77042381	276934087	Walt Disney	PG	Adventure
16	5/21/2010	Shrek Forever After	1.65e+08	238736787	756244673	Paramount Pictures	PG	Adventure
17	11/4/2016	Doctor Strange	1.65e+08	232641920	676486457	Walt Disney	PG-13	Action
18	11/7/2014	Big Hero 6	1.65e+08	222527828	652127828	Walt Disney	PG	Adventure
19	3/26/2010	How to Train Your Dragon	1.65e+08	217581232	494870992	Paramount Pictures	PG	Adventure
20	11/2/2012	Wreck-It Ralph	1.65e+08	189412677	496511521	Walt Disney	PG	Adventure

Univariate visualizations

Univariate visualizations describe one variable at a time.

For categorical variables, usually bar plots or pie charts are used.

For numerical variables, usually histograms, box plots, or density plots are used.

Summarizing categorical data with a table

table(movies$genre)

   Action Adventure    Comedy     Drama    Horror 
      573       481       813      1236       298 

Visualizing a categorical variable

The barplot function takes a named numeric vector as its first argument:

barplot(height = table(movies$genre))

Customizing base R plots

Base R plotting functions have many optional arguments allowing customization:

barplot(
  height = table(movies$genre), 
  main = "Distribution of Movie Genres in Tidy Tuesday dataset",
  col = "red"
)

Visualizing a categorical variable

The pie function takes a named numeric vector as its first argument:

pie(x = table(movies$genre))

Summarizing a numerical variable

summary(movies$worldwide_gross)

     Min.   1st Qu.    Median      Mean   3rd Qu.      Max. 
0.000e+00 1.062e+07 4.016e+07 9.412e+07 1.176e+08 1.305e+09 

Visualizing a numerical variable

hist(
  x = movies$worldwide_gross, 
  main = "Distribution of Worldwide Gross of Movies",
  col = "red"
)

What shape does this distribution have?

Visualizing a numerical variable

boxplot(
  x = movies$worldwide_gross, 
  main = "Distribution of Worldwide Gross",
  xlab = "Worldwide Gross ($)",
  col = "red",
  horizontal = T
)

Visualizing a numerical variable

boxplot(
  x = movies$production_budget, 
  main = "Distribution of Production Budget",
  xlab = "Budget ($)",
  col = "red",
  horizontal = T
)

Multivariate visualizations

How might we plot the relationship between a numerical variable and a categorical variable?

How about between two categorical variables? Between two numerical variables?

What if we want to visualize three variables at once?

Visualizing two numerical variables

The plot function can be used to obtain scatter plots for two numerical variables:

plot(
  x = movies$production_budget, 
  y = movies$worldwide_gross, 
  main = "Worldwide Gross vs. Production Budget",
  xlab = "Budget ($)",
  ylab = "Worldwide Gross ($)",
  col = "blue",
)

Visualizing two numerical variables

Visualizing a numerical variable and a categorical variable

The plot function can be used to obtain stacked boxplots for two numerical variables:

boxplot(worldwide_gross ~ genre,
        data = movies, 
        main = "Distribution of Worldwide Gross by Genre"
        xlab = "Worldwide Gross ($)",
        ylab = "",
        horizontal = T,
        las = 1)

Visualizing two categorical variables

table(movies$genre, movies$mpaa_rating)

           
              G  PG PG-13   R
  Action      1  34   225 286
  Adventure  67 309    83  14
  Comedy      6  79   328 367
  Drama      11 144   398 645
  Horror      0   7    58 202

Visualizing two categorical variables

barplot(table(movies$genre, movies$mpaa_rating),
        legend =  c("Action", "Adventure", "Comedy", "Drama", "Horror"))

Data visualization with ggplot2

ggplot2 is a fantastic package for making complex visualizations that many people find more intuitive to use.

library(ggplot2)

Plots are built sequentially using layers, so it’s easy to edit and fine-tune the plots you generate.

When using ggplot2, it is essential that your data are properly formatted If they are not, the results probably will not look like you expect.

Let’s work through how to build a plot layer by layer.

Data visualization with ggplot2

First, let’s initialize a plot. Use the data parameter to tell ggplot what data frame to use.

• It should be tidy data, in either a data.frame or tibble!

ggplot(data = movies)

Select variables using ggplot2

Next, we match variables to aesthetics within our plot. For example, let’s put Pollution Burden on the x-axis. We do so using the function aes() within the initial ggplot() call.

• Possible aesthetics include axes as well as characteristics such as color, size, shape

ggplot(data = movies,
       aes(x = production_budget)) 

Map aesthetics to shapes using ggplot2

Now ggplot knows what to plot, but it doesn’t know how to plot it yet. Let’s add a new layer that creates a histogram with geom_histogram().

• This is a new layer! We add layers using the + operator.

ggplot(data = movies,
       aes(x = production_budget)) +
  geom_histogram()

Customizing shapes

We can customize the shape within geom_histogram():

ggplot(data = movies,
       aes(x = production_budget)) +
  geom_histogram(fill = 'red')

Adding label and titles

We can finally add axis labels and a title using the + operator:

ggplot(data = movies,
       aes(x = production_budget)) +
  geom_histogram(fill = 'red') + 
  xlab("Budget ($)") +
  ggtitle("Distribution of Production Budgets")

Subsetting our data

Recall we can subset our data using the dplyr::filter() function from tidyverse. Note the use of the %in% operator here:

# remember you may need to run library(tidyverse)
select_genres <- c("Action", "Comedy")
# require county to be in SF Bay Area
action_comedy <- movies |>
  filter(genre %in% select_genres)
head(action_comedy)

# A tibble: 6 × 9
   ...1 release_date movie      production_budget domestic_gross worldwide_gross
  <dbl> <chr>        <chr>                  <dbl>          <dbl>           <dbl>
1     1 6/22/2007    Evan Almi…         175000000      100289690       174131329
2     2 7/28/1995    Waterworld         175000000       88246220       264246220
3     4 12/25/2013   47 Ronin           175000000       38362475       151716815
4     5 6/22/2018    Jurassic …         170000000      416769345      1304866322
5     6 8/1/2014     Guardians…         170000000      333172112       771051335
6     7 5/7/2010     Iron Man 2         170000000      312433331       621156389
# ℹ 3 more variables: distributor <chr>, mpaa_rating <chr>, genre <chr>

Scatter plots with ggplot2

We can easily add another variable/aesthetic pair and change the shape to get a multivariate visualization:

ggplot(data = action_comedy,
       aes(x = production_budget, y = worldwide_gross)) +
  geom_point()

Customizing our scatter plot

We can similarly customize our scatter plots

ggplot(data = action_comedy,
       aes(x = production_budget, y = worldwide_gross)) +
  geom_point(color = 'red', alpha = 0.5)  +
  xlab("Production Budget ($)") +
  ylab("Worldwide Gross ($)") +
  ggtitle("Worldwide Gross vs. Budget for \nAction/Comedy Movies")

Beyond two dimensions

We can add even more mappings between aesthetics and variables to explore multivariate relationships:

ggplot(data = action_comedy,
       aes(x = production_budget, y = worldwide_gross, color = genre)) +
  geom_point(alpha = 0.5)  +
  xlab("Production Budget ($)") +
  ylab("Worldwide Gross ($)") +
  ggtitle("Worldwide Gross vs. Budget for \nAction/Comedy Movies")

What to remember for ggplot2

The ggplot2 package is built around a “grammar of graphics”: a standard way of building plots layer by layer.

The three basic layers:

1. Data: a data frame with all of the variables of interest
2. Aesthetics: graphical dimensions like x, y, color, shape, and more.
3. Geometries: the specific markings used to illustrate your variables and aesthetics.

What to remember for ggplot2

Writing ggplot2 code can be confusing since most people are used to thinking about the geometry first.

One approach that may help:

1. Start by picking a geometry (ex. geom_point()).
2. Research the available aesthetics for that geometry (ex. x, y, color, shape, size). Not all aesthetics are available for all geometries.
3. Map variables to the dataset to the relevant aesthetics.

Beyond two dimensions

Be sure to avoid making your visualization overly complicated:

ggplot(data = movies,
       aes(x = production_budget, 
           y = worldwide_gross, 
           color = genre,
           shape = mpaa_rating,
           size = domestic_gross)) +
  geom_point(alpha = 0.5)  +
  xlab("Production Budget ($)") +
  ylab("Worldwide Gross ($)") +
  ggtitle("Worldwide Gross vs. Budget")

Beyond two dimensions

ggplot cheatsheet

• Axes: xlim(), ylim()
• Legends: within initial aes(), edit within theme() or guides()
• geom_point(), geom_line(), geom_histogram(), geom_bar(), geom_boxplot(), geom_text()
• facet_grid(), facet_wrap() for faceting
• labs() for labels
• theme_bw() to make things look nicer
• Graphical parameters: color for color, alpha for opacity, lwd/size for thickness, shape for shape, fill for interior color, …

Find more tips here!

General principles for data visualization

Avoid:

• 3D plots
• Deceptive axes
• Excessive labels
• Excessive/bad coloring
• Fancy shading/effects
• Bad variable/axis names
• Unreadable labels
• Overloaded with information

Strive for:

• Simple, clean graphics
• Neat and human readable text
• Appropriate data range (bar charts should always start from 0!)
• Consistent intervals
• Roughly ~6 colors or less
• Size figures appropriately

Sometimes breaking the rules is okay…

Charles Minard

Sometimes breaking the rules is okay…

W.E.B. DuBois

Sometimes breaking the rules is okay…

Florence Nightingale

Lab 04

In this lab, we’ll practice data visualization with base R and ggplot2.

Some misleading visualizations

Avoid: Pie charts? The proportions are completely off.

Some misleading visualizations

Avoid: Pie charts? Percentages don’t add up.

Some misleading visualizations

Avoid: 3d bar charts?!

Some misleading visualizations

Avoid: Bad use of color

Some misleading visualizations

Avoid: Bad axis, lack of information, 3D bar chart

Some misleading visualizations

Avoid: Four numbers displayed as a cluttered chart, terrible labels, bad axis range

Some misleading visualizations

Avoid: Deceptively flipped y-axis! (and excessive color)

Some misleading visualizations

Avoid: Two axes in a single plot, bad axis range

Some misleading visualizations

Avoid: Deceptive axis range (should start at 0)

Some misleading visualizations

Avoid: Inconsistent x-axis intervals