Deep dive into ggplot2 layers

Lecture 2

Dr. Mine Çetinkaya-Rundel

Duke University
STA 313 - Spring 2026

Warm up

Announcements

• Office hours + locations finalized at vizdata.org/course-team.html

• Don’t forget to complete the getting to know you survey by 8pm today

Setup

# load packages
library(tidyverse)
library(scales)
library(openintro)

# set theme for ggplot2
ggplot2::theme_set(ggplot2::theme_minimal(base_size = 14))

# set figure parameters for knitr
knitr::opts_chunk$set(
  fig.width = 7, # 7" width
  fig.asp = 0.618, # the golden ratio
  fig.retina = 3, # dpi multiplier for displaying HTML output on retina
  fig.align = "center", # center align figures
  dpi = 300 # higher dpi, sharper image
)

A/B testing

Data: Sale prices of houses in Duke Forest

• Data on houses that were sold in the Duke Forest neighborhood of Durham, NC around November 2020

• Scraped from Zillow

• Source: openintro::duke_forest

openintro::duke_forest

glimpse(duke_forest)

Rows: 98
Columns: 13
$ address    <chr> "1 Learned Pl, Durham, NC 27705", "1616 Pinecrest Rd, Durha…
$ price      <dbl> 1520000, 1030000, 420000, 680000, 428500, 456000, 1270000, …
$ bed        <dbl> 3, 5, 2, 4, 4, 3, 5, 4, 4, 3, 4, 4, 3, 5, 4, 5, 3, 4, 4, 3,…
$ bath       <dbl> 4.0, 4.0, 3.0, 3.0, 3.0, 3.0, 5.0, 3.0, 5.0, 2.0, 3.0, 3.0,…
$ area       <dbl> 6040, 4475, 1745, 2091, 1772, 1950, 3909, 2841, 3924, 2173,…
$ type       <chr> "Single Family", "Single Family", "Single Family", "Single …
$ year_built <dbl> 1972, 1969, 1959, 1961, 2020, 2014, 1968, 1973, 1972, 1964,…
$ heating    <chr> "Other, Gas", "Forced air, Gas", "Forced air, Gas", "Heat p…
$ cooling    <fct> central, central, central, central, central, central, centr…
$ parking    <chr> "0 spaces", "Carport, Covered", "Garage - Attached, Covered…
$ lot        <dbl> 0.97, 1.38, 0.51, 0.84, 0.16, 0.45, 0.94, 0.79, 0.53, 0.73,…
$ hoa        <chr> NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA,…
$ url        <chr> "https://www.zillow.com/homedetails/1-Learned-Pl-Durham-NC-…

A simple visualization

• Code
• Plot

ggplot(duke_forest, aes(x = area, y = price)) +
  geom_point(alpha = 0.7, size = 2) +
  geom_smooth(method = "lm", se = FALSE, linewidth = 0.7) +
  labs(
    x = "Area (square feet)",
    y = "Sale price (USD)",
    title = "Price and area of houses in Duke Forest"
  )

`geom_smooth()` using formula = 'y ~ x'

`geom_smooth()` using formula = 'y ~ x'

New variable: decade_built

duke_forest <- duke_forest |>
  mutate(decade_built = (year_built %/% 10) * 10)

duke_forest |>
  select(year_built, decade_built)

# A tibble: 98 × 2
   year_built decade_built
        <dbl>        <dbl>
 1       1972         1970
 2       1969         1960
 3       1959         1950
 4       1961         1960
 5       2020         2020
 6       2014         2010
 7       1968         1960
 8       1973         1970
 9       1972         1970
10       1964         1960
# ℹ 88 more rows

New variable: decade_built_cat

duke_forest <- duke_forest |>
  mutate(
    decade_built_cat = case_when(
      decade_built <= 1940 ~ "1940 or before",
      decade_built >= 1990 ~ "1990 or after",
      .default = as.character(decade_built)
    )
  )

duke_forest |>
  count(decade_built_cat)

# A tibble: 6 × 2
  decade_built_cat     n
  <chr>            <int>
1 1940 or before       8
2 1950                26
3 1960                32
4 1970                11
5 1980                13
6 1990 or after        8

A slightly more complex visualization

• Code
• Plot

ggplot(
  duke_forest,
  aes(x = area, y = price, color = decade_built_cat)
) +
  geom_point(alpha = 0.7, show.legend = FALSE) +
  geom_smooth(method = "lm", se = FALSE, linewidth = 0.5, show.legend = FALSE) +
  facet_wrap(~decade_built_cat) +
  labs(
    x = "Area (square feet)",
    y = "Sale price (USD)",
    color = "Decade built",
    title = "Price and area of houses in Duke Forest"
  )

`geom_smooth()` using formula = 'y ~ x'

`geom_smooth()` using formula = 'y ~ x'

A/B testing

In the next two slides, the same plots are created with different “cosmetic” choices. Examine the plots two given (Plot A and Plot B), and indicate your preference by voting for one of them in the Vote tab.

Test 1

• Plot A
• Plot B
• Vote

Test 2

• Plot A
• Plot B
• Vote

What makes figures bad?

Bad taste

Data-to-ink ratio

Tufte strongly recommends maximizing the data-to-ink ratio this in the Visual Display of Quantitative Information (Tufte, 1983).

Graphical excellence is the well-designed presentation of interesting data—a matter of substance, of statistics, and of design … [It] consists of complex ideas communicated with clarity, precision, and efficiency. … [It] is that which gives to the viewer the greatest number of ideas in the shortest time with the least ink in the smallest space … [It] is nearly always multivariate … And graphical excellence requires telling the truth about the data. (Tufte, 1983, p. 51).

Which of the plots has a higher data-to-ink ratio?

• Plot A
• Plot B
• Vote

A deeper look

at the plotting code

Summary statistics

mean_area_decade <- duke_forest |>
  group_by(decade_built_cat) |>
  summarize(mean_area = mean(area))

mean_area_decade

# A tibble: 6 × 2
  decade_built_cat mean_area
  <chr>                <dbl>
1 1940 or before       2072.
2 1950                 2545.
3 1960                 2873.
4 1970                 3413.
5 1980                 2889.
6 1990 or after        2822.

Barplot

• Code
• Plot

ggplot(
  mean_area_decade,
  aes(y = decade_built_cat, x = mean_area)
) +
  geom_col() +
  labs(
    x = "Mean area (square feet)",
    y = "Decade built",
    title = "Mean area of houses in Duke Forest, by decade built"
  )

Scaterplot

• Code
• Plot

ggplot(
  mean_area_decade,
  aes(y = decade_built_cat, x = mean_area)
) +
  geom_point(size = 4) +
  labs(
    x = "Mean area (square feet)",
    y = "Decade built",
    title = "Mean area of houses in Duke Forest, by decade built"
  )

Lollipop chart – a happy medium?

Application exercise

• Go to the course GitHub organization: https://github.com/vizdata-s24

• Clone the repo called ae-01-YOUR-GITHUB-NAME and work on the exercise.

• Once you’re done, commit and push your changes to GitHub.

• Label your cell(s) and pay attention to code style and formatting!

Bad data

• Original
• Improved

Healy, Data Visualization: A practical introduction. Chapter 1. Figures 1.8 and 1.9.

Bad perception

Healy, Data Visualization: A practical introduction. Chapter 1. Figure 1.12.

Aesthetic mappings in ggplot2

A second look: lollipop chart

• Plot
• Code

ggplot(
  mean_area_decade,
  aes(y = decade_built_cat, x = mean_area)
) +
  geom_point(size = 4) +
  geom_segment(aes(
    x = 0,
    xend = mean_area,
    y = decade_built_cat,
    yend = decade_built_cat
  )) +
  labs(
    x = "Mean area (square feet)",
    y = "Decade built",
    title = "Mean area of houses in Duke Forest, by decade built"
  )

Activity: Spot the differences I

Can you spot the differences between the code here and the one provided in the previous slide? Are there any differences in the resulting plot? Work in a pair (or group) to answer.

• Plot
• Code

ggplot(
  mean_area_decade,
  aes(y = decade_built_cat, x = mean_area)
) +
  geom_point(size = 4) +
  geom_segment(aes(
    xend = 0,
    yend = decade_built_cat
  )) +
  labs(
    x = "Mean area (square feet)",
    y = "Decade built",
    title = "Mean area of houses in Duke Forest, by decade built"
  )

Global vs. layer-specific aesthetics

• Aesthetic mappings can be supplied in the initial ggplot() call, in individual layers, or in some combination of both.

• Within each layer, you can add, override, or remove mappings.

• If you only have one layer in the plot, the way you specify aesthetics doesn’t make any difference. However, the distinction is important when you start adding additional layers.

Activity: Spot the differences II

Do you expect the following plots to be the same or different? If different, how? Discuss in a pair (or group) without running the code and sketch the resulting plots based on what you think the code will produce.

# Plot A
ggplot(duke_forest, aes(x = area, y = price)) +
  geom_point(aes(color = decade_built_cat))

# Plot B
ggplot(duke_forest, aes(x = area, y = price)) +
  geom_point(color = "blue")

# Plot C
ggplot(duke_forest, aes(x = area, y = price)) +
  geom_point(color = "#a493ba")

Wrap up

Think back to all the plots you saw in the lecture, without flipping back through the slides. Which plot first comes to mind?

Geoms

Geoms

• Geometric objects, or geoms for short, perform the actual rendering of the layer, controlling the type of plot that you create

• You can think of them as “the geometric shape used to represent the data”

One variable

• Discrete:

  • geom_bar(): display distribution of discrete variable.

• Continuous

  • geom_histogram(): bin and count continuous variable, display with bars

  • geom_density(): smoothed density estimate

  • geom_dotplot(): stack individual points into a dot plot

  • geom_freqpoly(): bin and count continuous variable, display with lines

Aside

Always use “typewriter text” (monospace font) when writing function names, and follow with (), e.g.,

• geom_freqpoly()

• mean()

• lm()

geom_dotplot()

What does each point represent? How are their locations determined? What do the x and y axes represent?

ggplot(duke_forest, aes(x = price)) +
  geom_dotplot(binwidth = 50000, dotsize = 0.7)

Comparing across groups

Which of the following allows for easier comparison across groups?

• Histogram
• Frequency polygon

ggplot(duke_forest, aes(x = price, fill = decade_built_cat)) +
  geom_histogram(binwidth = 100000)

ggplot(duke_forest, aes(x = price, color = decade_built_cat)) +
  geom_freqpoly(binwidth = 100000, linewidth = 1)

Two variables - both continuous

• geom_point(): scatterplot

• geom_quantile(): smoothed quantile regression

• geom_rug(): marginal rug plots

• geom_smooth(): smoothed line of best fit

• geom_text(): text labels

Two variables - show density

• geom_bin2d(): bin into rectangles and count

• geom_density2d(): smoothed 2d density estimate

• geom_hex(): bin into hexagons and count

geom_hex()

Not very helpful for 98 observations:

ggplot(duke_forest, aes(x = area, y = price)) +
  geom_hex()

geom_hex()

More helpful for 53940 observations:

ggplot(diamonds, aes(x = carat, y = price)) +
  geom_hex()

geom_hex()

(Maybe) even more helpful on the (natural) log scale:

ggplot(diamonds, aes(x = carat, y = price)) +
  geom_hex() +
  scale_fill_gradient(trans = "log")

geom_hex() and warnings

• Requires installing the hexbin package separately!

install.packages("hexbin")

• Otherwise you might see

Warning: Computation failed in `stat_binhex()`

Two variables

• At least one discrete
  • geom_count(): count number of point at distinct locations
  • geom_jitter(): randomly jitter overlapping points
• One continuous, one discrete
  • geom_col(): a bar chart of pre-computed summaries
  • geom_boxplot(): boxplots
  • geom_violin(): show density of values in each group

geom_jitter()

How are the following three plots different?

• Plot A
• Plot B
• Plot C

ggplot(duke_forest, aes(x = bed, y = price)) +
  geom_point()

ggplot(duke_forest, aes(x = bed, y = price)) +
  geom_jitter()

ggplot(duke_forest, aes(x = bed, y = price)) +
  geom_jitter()

geom_jitter() and set.seed()

• Plot A
• Plot B

set.seed(1234)

ggplot(duke_forest, aes(x = bed, y = price)) +
  geom_jitter()

set.seed(1234)

ggplot(duke_forest, aes(x = bed, y = price)) +
  geom_jitter()

Two variables

• One time, one continuous
  • geom_area(): area plot
  • geom_line(): line plot
  • geom_step(): step plot
• Display uncertainty:
  • geom_crossbar(): vertical bar with center
  • geom_errorbar(): error bars
  • geom_linerange(): vertical line
  • geom_pointrange(): vertical line with center
• Spatial
  • geom_sf(): for map data (more on this later…)

Average price per year built

mean_price_year <- duke_forest |>
  group_by(year_built) |>
  summarize(
    n = n(),
    mean_price = mean(price),
    sd_price = sd(price)
  )

mean_price_year

# A tibble: 44 × 4
   year_built     n mean_price sd_price
        <dbl> <int>      <dbl>    <dbl>
 1       1923     1     285000      NA 
 2       1934     1     600000      NA 
 3       1938     1     265000      NA 
 4       1940     1     105000      NA 
 5       1941     2     432500   28284.
 6       1945     2     525000  530330.
 7       1951     2     567500  258094.
 8       1952     2     531250  469165.
 9       1953     2     575000   35355.
10       1954     4     600000   33912.
# ℹ 34 more rows

geom_line()

ggplot(mean_price_year, aes(x = year_built, y = mean_price)) +
  geom_line()

geom_area()

ggplot(mean_price_year, aes(x = year_built, y = mean_price)) +
  geom_area()

geom_step()

ggplot(mean_price_year, aes(x = year_built, y = mean_price)) +
  geom_step()