More on ggplot

2025-04-01

Rick Gilmore

Overview

Announcements

• Exercise 06 due Thursday, April 3.
• Final project presentation survey

DataCamp

Last time

data_frame |>
  # Manipulations to the data
  ggplot() +
  aes(x = , y = , fill = ) + # label aesthetic mappings
  # Add geom(s)
  geom_* +
  # Format legends, axes, add title, etc.
  ggtitle("My awesome figure")

Today’s topics

• A future of data science
• Optimism and data visualization
• More ggplot2
• Work session

A future of data science

Downey (2024)

Your turn

• Is Downey right?

Optimism and data visualization

Hans Rosling

Rosling (n.d.)

Gilmore’s view

• Things are much better than they seem
• Especially if we look at the data
• We need more data, more widely & publicly available
• Plus skilled interpreters (YOU!)
• Knowledge and skepticism (+ action) are antidotes to pessimism

More ggplot2

“National Survey of Family Growth” (2024)

Gathering data

• Visit https://www.cdc.gov/nchs/nsfg/nsfg-2022-2023-puf.htm
• Download data (https://ftp.cdc.gov/pub/Health_Statistics/NCHS/NSFG/NSFG-2022-2023-FemPregPUFData.zip), unzip
• Copy to computer
  • Data: src/include/csv
  • Codebook: src/include/pdf

Importing data

library(readr)
library(here)

path2data <- file.path(here::here(), "src", "include", "csv")

nsfg_22_23 <- read_csv(
  file.path(path2data, "NSFG_2022_2023_FemPregPUFData.csv"),
  show_col_types = FALSE,
  guess_max = Inf
)

Examining data

str(nsfg_22_23)

spc_tbl_ [8,247 × 111] (S3: spec_tbl_df/tbl_df/tbl/data.frame)
 $ CaseID      : num [1:8247] 96071 96072 96072 96072 96072 ...
 $ PREGORDR    : num [1:8247] 1 4 3 2 1 2 1 1 2 1 ...
 $ FTFMODE     : num [1:8247] 2 2 2 2 2 2 2 2 2 2 ...
 $ BORNALIV    : num [1:8247] 1 1 1 1 1 1 NA 1 1 1 ...
 $ RECNT5YRPRG : num [1:8247] 1 5 5 5 5 1 1 5 5 5 ...
 $ BABYSEX     : num [1:8247] 1 2 2 1 1 1 NA 2 2 1 ...
 $ BIRTHWGT    : num [1:8247] 1 1 1 1 1 1 NA 2 1 1 ...
 $ PAYBIRTH1   : num [1:8247] 1 NA NA NA NA 1 NA NA NA NA ...
 $ PAYBIRTH2   : num [1:8247] NA NA NA NA NA NA NA NA NA NA ...
 $ PAYBIRTH3   : num [1:8247] NA NA NA NA NA NA NA NA NA NA ...
 $ AGEFATHER   : num [1:8247] 5 3 2 2 1 2 2 5 5 3 ...
 $ KNEWPREG    : num [1:8247] 1 NA NA NA NA 1 1 NA NA NA ...
 $ PRIORSMK    : num [1:8247] 0 NA NA NA NA 4 4 NA NA NA ...
 $ POSTSMKS    : num [1:8247] 5 NA NA NA NA 1 5 NA NA NA ...
 $ GETPRENA    : num [1:8247] 1 NA NA NA NA 1 5 NA NA NA ...
 $ BGNPRENA    : num [1:8247] 1 NA NA NA NA 1 NA NA NA NA ...
 $ KIDAGE      : num [1:8247] 2 8 8 8 8 2 NA 6 6 8 ...
 $ FRSTEATD    : num [1:8247] 0 NA NA NA NA NA NA NA NA NA ...
 $ RECPREG5YR  : num [1:8247] 1 5 5 5 5 1 1 5 5 5 ...
 $ RECPREG3YR  : num [1:8247] 1 5 5 5 5 1 1 5 5 5 ...
 $ EVUSEINT    : num [1:8247] NA 1 1 1 5 NA NA 1 1 1 ...
 $ STOPDUSE    : num [1:8247] NA 1 1 1 NA NA NA 1 1 1 ...
 $ WHYSTOPD    : num [1:8247] NA 5 5 5 NA NA NA 5 1 1 ...
 $ WHATMETH1   : num [1:8247] NA NA NA NA NA NA NA NA NA NA ...
 $ WHATMETH2   : num [1:8247] NA NA NA NA NA NA NA NA NA NA ...
 $ WHATMETH3   : num [1:8247] NA NA NA NA NA NA NA NA NA NA ...
 $ WHATMETH4   : num [1:8247] NA NA NA NA NA NA NA NA NA NA ...
 $ RESNOUSE    : num [1:8247] 1 NA NA NA 5 1 1 NA NA NA ...
 $ WANTBOLD    : num [1:8247] NA 5 5 5 5 NA NA 1 NA NA ...
 $ PROBBABE    : num [1:8247] NA NA NA NA NA NA NA NA NA NA ...
 $ TIMINGOK    : num [1:8247] 3 NA NA NA NA 2 2 2 2 2 ...
 $ TOOSOON_N   : num [1:8247] NA NA NA NA NA NA NA NA NA NA ...
 $ TOOSOON_P   : num [1:8247] NA NA NA NA NA NA NA NA NA NA ...
 $ LATERNUM    : num [1:8247] 2 NA NA NA NA NA NA NA NA NA ...
 $ LATERMY     : num [1:8247] 2 NA NA NA NA NA NA NA NA NA ...
 $ WTHPART1    : num [1:8247] 1 NA NA NA NA 1 NA NA NA NA ...
 $ WTHPART2    : num [1:8247] NA NA NA NA NA NA NA NA NA NA ...
 $ FEELINPG    : num [1:8247] 10 NA NA NA NA 7 6 NA NA NA ...
 $ HPWNOLD     : num [1:8247] 1 5 5 5 5 1 1 6 1 1 ...
 $ TIMOKHP     : num [1:8247] 3 NA NA NA NA 2 2 NA 2 2 ...
 $ COHPBEG     : num [1:8247] NA 1 1 1 3 2 2 3 1 1 ...
 $ COHPEND     : num [1:8247] NA 1 1 1 3 2 2 3 1 1 ...
 $ TELLFATH    : num [1:8247] NA NA NA NA NA NA NA NA NA NA ...
 $ WHENTELL    : num [1:8247] NA NA NA NA NA NA NA NA NA NA ...
 $ TRYSCALE    : num [1:8247] 10 NA NA NA NA 10 5 NA NA NA ...
 $ WANTSCAL    : num [1:8247] 10 NA NA NA NA 10 5 NA NA NA ...
 $ WHYPRG1     : num [1:8247] NA NA NA NA NA NA NA NA NA NA ...
 $ WHYPRG2     : num [1:8247] NA NA NA NA NA NA NA NA NA NA ...
 $ WHYNOUSE1   : num [1:8247] NA NA NA NA NA NA NA NA NA NA ...
 $ WHYNOUSE2   : num [1:8247] NA NA NA NA NA NA NA NA NA NA ...
 $ WHYNOUSE3   : num [1:8247] NA NA NA NA NA NA NA NA NA NA ...
 $ WHYNOUSE4   : num [1:8247] NA NA NA NA NA NA NA NA NA NA ...
 $ MAINOUSE    : num [1:8247] NA NA NA NA NA NA NA NA NA NA ...
 $ OUTCOME     : num [1:8247] 1 1 1 1 1 1 4 1 1 1 ...
 $ BIRTHORD    : num [1:8247] 1 4 3 2 1 1 NA 1 2 1 ...
 $ DATEND      : num [1:8247] 2021 1999 1993 1991 1989 ...
 $ agepreg     : num [1:8247] 38 25 20 18 16 23 22 29 34 21 ...
 $ GEST_LB     : num [1:8247] 3 3 3 3 3 3 NA 3 3 3 ...
 $ GEST_OTHR   : num [1:8247] NA NA NA NA NA NA 1 NA NA NA ...
 $ GestImp     : num [1:8247] 2 2 2 2 2 2 2 2 2 2 ...
 $ DATECON     : num [1:8247] 2020 1998 1993 1991 1989 ...
 $ agecon      : num [1:8247] 37 25 19 17 15 22 22 28 33 20 ...
 $ pmarpreg    : num [1:8247] 2 2 2 2 1 1 1 1 2 2 ...
 $ BFEEDWKS    : num [1:8247] 3 NA NA NA NA 0 NA NA NA NA ...
 $ RMAROUT3    : num [1:8247] 1 1 1 1 3 2 2 3 1 1 ...
 $ RMARCON3    : num [1:8247] 1 1 1 1 3 2 2 3 1 1 ...
 $ NEWWANTR    : num [1:8247] 1 6 6 6 6 2 2 2 2 2 ...
 $ wantresp    : num [1:8247] 1 5 5 5 5 2 2 2 2 2 ...
 $ WANTPART    : num [1:8247] 1 5 5 5 5 2 2 6 2 2 ...
 $ AGER        : num [1:8247] 40 49 49 49 49 25 25 37 44 44 ...
 $ AGESCRN     : num [1:8247] 40 49 49 49 49 25 25 37 44 44 ...
 $ FMARITAL    : num [1:8247] 1 3 3 3 3 5 5 5 3 3 ...
 $ RMARITAL    : num [1:8247] 1 4 4 4 4 2 2 6 4 4 ...
 $ HIEDUC      : num [1:8247] 10 9 9 9 9 5 5 6 8 8 ...
 $ HISPANIC    : num [1:8247] 2 2 2 2 2 2 2 2 2 2 ...
 $ HISPRACE2   : num [1:8247] 2 3 3 3 3 4 4 2 2 2 ...
 $ BRNOUT      : num [1:8247] 5 5 5 5 5 1 1 5 5 5 ...
 $ PREGNUM     : num [1:8247] 1 4 4 4 4 2 2 1 2 2 ...
 $ PARITY      : num [1:8247] 1 4 4 4 4 1 1 1 2 2 ...
 $ CURR_INS    : num [1:8247] 1 4 4 4 4 3 3 2 1 1 ...
 $ POVERTY     : num [1:8247] 417 444 444 444 444 97 97 140 444 444 ...
 $ LABORFOR    : num [1:8247] 1 6 6 6 6 6 6 4 1 1 ...
 $ RELIGION    : num [1:8247] 1 1 1 1 1 1 1 3 3 3 ...
 $ METRO       : num [1:8247] 2 1 1 1 1 3 3 3 1 1 ...
 $ OUTCOME_I   : num [1:8247] 0 0 0 0 0 0 0 0 0 0 ...
 $ DATEND_I    : num [1:8247] 0 0 0 0 0 0 0 0 0 0 ...
 $ PRGLNGTH_I  : num [1:8247] 0 0 0 0 0 0 0 0 0 0 ...
 $ DATECON_I   : num [1:8247] 0 0 0 0 0 0 0 0 0 0 ...
 $ BFEEDWKS_I  : num [1:8247] 0 0 0 0 0 0 0 0 0 0 ...
 $ RMAROUT3_I  : num [1:8247] 0 0 0 0 0 0 0 0 0 0 ...
 $ RMARCON3_I  : num [1:8247] 0 0 0 0 0 0 0 0 0 0 ...
 $ NEWWANTR_I  : num [1:8247] 0 0 0 0 0 0 0 0 0 0 ...
 $ WANTPART_I  : num [1:8247] 0 0 0 0 0 0 0 0 0 0 ...
 $ HIEDUC_I    : num [1:8247] 0 0 0 0 0 0 0 0 0 0 ...
 $ HISPANIC_I  : num [1:8247] 0 0 0 0 0 0 0 0 0 0 ...
 $ HISPRACE2_I : num [1:8247] 0 0 0 0 0 0 0 0 0 0 ...
 $ PREGNUM_I   : num [1:8247] 0 0 0 0 0 0 0 0 0 0 ...
 $ CURR_INS_I  : num [1:8247] 0 0 0 0 0 0 0 0 0 0 ...
 $ POVERTY_I   : num [1:8247] 0 0 0 0 0 0 0 0 0 0 ...
  [list output truncated]
 - attr(*, "spec")=
  .. cols(
  ..   CaseID = col_double(),
  ..   PREGORDR = col_double(),
  ..   FTFMODE = col_double(),
  ..   BORNALIV = col_double(),
  ..   RECNT5YRPRG = col_double(),
  ..   BABYSEX = col_double(),
  ..   BIRTHWGT = col_double(),
  ..   PAYBIRTH1 = col_double(),
  ..   PAYBIRTH2 = col_double(),
  ..   PAYBIRTH3 = col_double(),
  ..   AGEFATHER = col_double(),
  ..   KNEWPREG = col_double(),
  ..   PRIORSMK = col_double(),
  ..   POSTSMKS = col_double(),
  ..   GETPRENA = col_double(),
  ..   BGNPRENA = col_double(),
  ..   KIDAGE = col_double(),
  ..   FRSTEATD = col_double(),
  ..   RECPREG5YR = col_double(),
  ..   RECPREG3YR = col_double(),
  ..   EVUSEINT = col_double(),
  ..   STOPDUSE = col_double(),
  ..   WHYSTOPD = col_double(),
  ..   WHATMETH1 = col_double(),
  ..   WHATMETH2 = col_double(),
  ..   WHATMETH3 = col_double(),
  ..   WHATMETH4 = col_double(),
  ..   RESNOUSE = col_double(),
  ..   WANTBOLD = col_double(),
  ..   PROBBABE = col_double(),
  ..   TIMINGOK = col_double(),
  ..   TOOSOON_N = col_double(),
  ..   TOOSOON_P = col_double(),
  ..   LATERNUM = col_double(),
  ..   LATERMY = col_double(),
  ..   WTHPART1 = col_double(),
  ..   WTHPART2 = col_double(),
  ..   FEELINPG = col_double(),
  ..   HPWNOLD = col_double(),
  ..   TIMOKHP = col_double(),
  ..   COHPBEG = col_double(),
  ..   COHPEND = col_double(),
  ..   TELLFATH = col_double(),
  ..   WHENTELL = col_double(),
  ..   TRYSCALE = col_double(),
  ..   WANTSCAL = col_double(),
  ..   WHYPRG1 = col_double(),
  ..   WHYPRG2 = col_double(),
  ..   WHYNOUSE1 = col_double(),
  ..   WHYNOUSE2 = col_double(),
  ..   WHYNOUSE3 = col_double(),
  ..   WHYNOUSE4 = col_double(),
  ..   MAINOUSE = col_double(),
  ..   OUTCOME = col_double(),
  ..   BIRTHORD = col_double(),
  ..   DATEND = col_double(),
  ..   agepreg = col_double(),
  ..   GEST_LB = col_double(),
  ..   GEST_OTHR = col_double(),
  ..   GestImp = col_double(),
  ..   DATECON = col_double(),
  ..   agecon = col_double(),
  ..   pmarpreg = col_double(),
  ..   BFEEDWKS = col_double(),
  ..   RMAROUT3 = col_double(),
  ..   RMARCON3 = col_double(),
  ..   NEWWANTR = col_double(),
  ..   wantresp = col_double(),
  ..   WANTPART = col_double(),
  ..   AGER = col_double(),
  ..   AGESCRN = col_double(),
  ..   FMARITAL = col_double(),
  ..   RMARITAL = col_double(),
  ..   HIEDUC = col_double(),
  ..   HISPANIC = col_double(),
  ..   HISPRACE2 = col_double(),
  ..   BRNOUT = col_double(),
  ..   PREGNUM = col_double(),
  ..   PARITY = col_double(),
  ..   CURR_INS = col_double(),
  ..   POVERTY = col_double(),
  ..   LABORFOR = col_double(),
  ..   RELIGION = col_double(),
  ..   METRO = col_double(),
  ..   OUTCOME_I = col_double(),
  ..   DATEND_I = col_double(),
  ..   PRGLNGTH_I = col_double(),
  ..   DATECON_I = col_double(),
  ..   BFEEDWKS_I = col_double(),
  ..   RMAROUT3_I = col_double(),
  ..   RMARCON3_I = col_double(),
  ..   NEWWANTR_I = col_double(),
  ..   WANTPART_I = col_double(),
  ..   HIEDUC_I = col_double(),
  ..   HISPANIC_I = col_double(),
  ..   HISPRACE2_I = col_double(),
  ..   PREGNUM_I = col_double(),
  ..   CURR_INS_I = col_double(),
  ..   POVERTY_I = col_double(),
  ..   LABORFOR_I = col_double(),
  ..   RELIGION_I = col_double(),
  ..   WGT2022_2023 = col_double(),
  ..   VEST = col_double(),
  ..   VECL = col_double(),
  ..   CMINTVW = col_double(),
  ..   CMLSTYR = col_double(),
  ..   CMJAN3YR = col_double(),
  ..   CMJAN4YR = col_double(),
  ..   CMJAN5YR = col_double(),
  ..   YEAR = col_double(),
  ..   QUARTER = col_double()
  .. )
 - attr(*, "problems")=<externalptr> 

Help from the codebook

include/pdf/2022-2023-NSFG-FemRespPUFCodebook.pdf

Respondent age (AGER)

library(ggplot2)
nsfg_22_23 |>
  ggplot() +
  aes(x = AGER) +
  geom_histogram()

Respondent age (AGER)

library(ggplot2)
nsfg_22_23 |>
  ggplot() +
  aes(x = AGER) +
  geom_histogram()

Respondent age (AGER)

Figure 1: Histogram of NSFG 22-23 ‘AGER’ variable

Respondent age (AGER): Violin

nsfg_22_23 |>
  ggplot() +
  aes(x = AGER) +
  geom_violin()

Respondent age (AGER): Violin

Figure 2: Error message when I tried to execute this code.

Need help?

• Overall: https://ggplot2.tidyverse.org
• On geoms? https://ggplot2.tidyverse.org/reference/index.html#geoms

Respondent age (AGER): Violin

nsfg_22_23 |>
  ggplot() +
  aes(x = "", y = AGER) +
  geom_violin()

Respondent age (AGER): Violin

nsfg_22_23 |>
  ggplot() +
  aes(x = "", y = AGER) +
  geom_violin()

Respondent age (AGER): Violin

Figure 3: Violin plot of NSFG 22-23 ‘AGER’ variable

Respondent age (AGER): Boxplot

nsfg_22_23 |>
  ggplot() +
  aes(x = "", y = AGER) +
  geom_boxplot()

Respondent age (AGER): Boxplot

nsfg_22_23 |>
  ggplot() +
  aes(x = "", y = AGER) +
  geom_boxplot()

Respondent age (AGER): Boxplot

Figure 4: Boxplot of NSFG 22-23 ‘AGER’ variable

Respondent age (AGER): Composite

p_violin <- nsfg_22_23 |>
  ggplot() +
  aes(x = "", y = AGER) +
  geom_violin()
  
p_violin

Respondent age (AGER): Composite

p_violin <- nsfg_22_23 |>
  ggplot() +
  aes(x = "", y = AGER) +
  geom_violin()

p_violin

Respondent age (AGER): Composite

Figure 5: Distribution of NSFG 22-23 ‘AGER’ variable

Respondent age (AGER): Composite

p_violin_box <- p_violin +
  geom_boxplot()
  
p_violin_box

Respondent age (AGER): Composite

p_violin_box <- p_violin +
  geom_boxplot()
  
p_violin_box

Respondent age (AGER): Composite

Figure 6: Distribution of NSFG 22-23 ‘AGER’ variable

Respondent age (AGER): Composite

p_violin_box_points <- p_violin_box +
  geom_jitter(width = .1, height = 0)
  
p_violin_box_points

Respondent age (AGER): Composite

p_violin_box_points <- p_violin_box +
  geom_jitter(width = .15, height = 0, alpha=.1)
  
p_violin_box_points

Respondent age (AGER): Composite

Figure 7: Distribution of NSFG 22-23 ‘AGER’ variable

Respondent age (AGER): Composite

nsfg_22_23 |>
  ggplot() +
  aes(x = "", y = AGER) +
  geom_violin() +
  geom_boxplot() +
  geom_jitter(width = .15, height = 0, alpha=.1)

Respondent age (AGER): Composite

nsfg_22_23 |>
  ggplot() +
  aes(x = "", y = AGER) +
  geom_violin() +
  geom_boxplot() +
  geom_jitter(width = .15, height = 0, alpha=.1)

Respondent age (AGER): Composite

Figure 8: Distribution of NSFG 22-23 ‘AGER’ variable

Respondent age (AGER): Composite

nsfg_22_23 |>
  ggplot() +
  aes(x = "", y = AGER) +
  geom_boxplot() +
  geom_violin() +
  geom_jitter(width = .15, height = 0, alpha=.1)

• Boxplot first, then violin

Respondent age (AGER): Composite

nsfg_22_23 |>
  ggplot() +
  aes(x = "", y = AGER) +
  geom_boxplot() +
  geom_violin() +
  geom_jitter(width = .15, height = 0, alpha=.1)

Respondent age (AGER): Composite

Figure 9: Distribution of NSFG 22-23 ‘AGER’ variable

Order matters!

When adding multiple layers of geoms, think about what order makes the most sense.

Here, the best order is violin, then boxplot, then points.

Iterative enhancement

nsfg_22_23 |>
  ggplot() +
  aes(x = "", y = AGER) +
  geom_violin() +
  geom_boxplot() +
  geom_jitter(width = .15, height = 0, alpha=.1) +
  ylab("Respondent age (years)") +
  xlab("")

Iterative enhancement

nsfg_22_23 |>
  ggplot() +
  aes(x = "", y = AGER) +
  geom_violin() +
  geom_boxplot() +
  geom_jitter(width = .15, height = 0, alpha=.1) +
  ylab("Respondent age (years)") +
  xlab("")

Iterative enhancement

Figure 10: Distribution of NSFG 22-23 ‘AGER’ variable

New twist!

nsfg_22_23 |>
  ggplot() +
  aes(x = "", y = AGER) +
  geom_violin() +
  geom_boxplot() +
  geom_jitter(width = .15, height = 0, alpha=.1) +
  ylab("Respondent age (years)") +
  xlab("") +
  facet_wrap(~RELIGION)

New twist!

nsfg_22_23 |>
  ggplot() +
  aes(x = "", y = AGER) +
  geom_violin() +
  geom_boxplot() +
  geom_jitter(width = .15, height = 0, alpha=.1) +
  ylab("Respondent age (years)") +
  xlab("") +
  facet_wrap(~RELIGION)

New twist!

Figure 11: Distribution of NSFG 22-23 ‘AGER’ variable by ‘RELIGION’

What do the numbers mean?

• Back to the codebook

Figure 12: RELIGION code from NSFG codebook

New twist!

library(dplyr)
nsfg_22_23 |>
  mutate(religion = case_match(RELIGION,
                               1 ~ "none",
                               2 ~ "catholic",
                               3 ~ "protestant",
                               4 ~ "other")) |>
  ggplot() +
  aes(x = "", y = AGER) +
  geom_violin() +
  geom_boxplot() +
  #geom_jitter(width = .15, height = 0, alpha=.1) +
  ylab("Respondent age (years)") +
  xlab("") +
  facet_wrap(~religion)

• Give numbers meaningful labels with dplyr::mutate() and dplyr::case_match().

New twist!

library(dplyr)
nsfg_22_23 |>
  mutate(religion = case_match(RELIGION,
                               1 ~ "none",
                               2 ~ "catholic",
                               3 ~ "protestant",
                               4 ~ "other")) |>
  ggplot() +
  aes(x = "", y = AGER) +
  geom_violin() +
  geom_boxplot() +
  #geom_jitter(width = .15, height = 0, alpha=.1) +
  ylab("Respondent age (years)") +
  xlab("") +
  facet_wrap(~religion)

New twist!

Figure 13: Distribution of NSFG 22-23 ‘AGER’ variable by religion

Recommended design principles

• Work iteratively (in small steps)
• Keep prior versions
  • Store figure code in *.R or *.qmd files
• No ‘secret sauce’
• The process is the product

Next time…

Python logo

Work session

Resources

References

Downey, A. (2024). A future of data science. Retrieved March 31, 2025, from https://www.youtube.com/watch?v=YKMZIzYBgTk&list=PL9HYL-VRX0oSFkdF4fJeY63eGDvgofcbn

National Survey of Family Growth. (2024, December 10). Retrieved March 31, 2025, from https://www.cdc.gov/nchs/nsfg/index.htm

Rosling, H. (n.d.). Hans rosling’s 200 countries, 200 years, 4 minutes - the joy of stats - BBC. Retrieved from https://www.youtube.com/watch?v=jbkSRLYSojo&t=80s