learning log example

This week I have been playing with tables

Jenny Richmond true
02-13-2021

This post is an example of the kind of log that you should post to RPubs each week. The learning log is an opportunity to reflect on what you have learned each week and to think about what the next steps in your coding journey are. It should answer the following questions…

  1. What were your coding goals this week?
  2. How did you go achieving your goals (challenges/successes)?
  3. What the next steps are in your coding journey?

… include examples of plots/code that you have been working on.

Jenny’s learning log I

library(tidyverse)
library(kableExtra)
library(gt)
library(DT)

This week the Tidy Tuesday dataset is about income disparities in the US. The student debt dataset looks particularly interesting.

debt <- read_csv("debt.csv")

glimpse(debt)

Rows: 30
Columns: 4
$ year          <dbl> 2016, 2016, 2016, 2013, 2013, 2013, 2010, 201…
$ race          <chr> "White", "Black", "Hispanic", "White", "Black…
$ loan_debt     <dbl> 11108.410, 14224.770, 7493.999, 8363.605, 103…
$ loan_debt_pct <dbl> 0.3367511, 0.4183588, 0.2189689, 0.2845555, 0…

Looks like year is numeric, make it a factor so that it works better in the plot.

debt$year <- as_factor(debt$year)

Looks like this df average family student loan debt for aged 25-55, by race and year normalized to 2016 dollars. Lets plot debt over time by race.

debt %>%
  ggplot(aes(x = year, y = loan_debt, colour = race, group = race)) +
  geom_point() +
  geom_line() +
  theme_classic() +
  scale_y_continuous(limits = c(0,15000)) +
  labs(title = "Average family student loan debt by race and year", 
       y = "Loan debt ($2016)", 
       x = "Year")

The goal

OK plotting is great, but I am trying to learn about tables this week. My goal is to practice making tables, so lets start by averaging this data into something that might be table worthy.

This chunk groups the debt data by year (averaging across race) and summarising the mean debt levels.

summary <- debt %>%
  group_by(year) %>%
  summarise(meandebt = mean(loan_debt))

summary

# A tibble: 10 x 2
   year  meandebt
 * <fct>    <dbl>
 1 1989     1053.
 2 1992     1014.
 3 1995     1735.
 4 1998     1959.
 5 2001     2311.
 6 2004     3225.
 7 2007     4793.
 8 2010     6880.
 9 2013     7281.
10 2016    10942.

I know that there are several packages that are useful for making tables in R. My goal is to try them out and work out what the pros/cons are

  1. kableExtra

  2. gt

  3. DT

Option 1 kableExtra

Pipe your dataframe into the kbl() function and you get a basic html table. Add kable styling to get a bootstrap theme.

summary %>%
  kbl()
year meandebt
1989 1052.853
1992 1013.949
1995 1735.111
1998 1959.001
2001 2311.100
2004 3224.881
2007 4793.357
2010 6880.418
2013 7281.225
2016 10942.393 
summary %>%
  kbl() %>%
  kable_styling()
year meandebt
1989 1052.853
1992 1013.949
1995 1735.111
1998 1959.001
2001 2311.100
2004 3224.881
2007 4793.357
2010 6880.418
2013 7281.225
2016 10942.393

It is weird that in each of these options, when I run the Rmd chunk it doesn’t display the contents of the table. I just get a white box. But… when I knit the document, they render just fine. Not sure what is going on with that. I also don’t really know what kable_styling() does; the only difference in the knitting document is that it appears centred on the page. There must be more to it….

Option 2 gt

The gt package makes great tables (apparently).

summary %>%
  gt()
year meandebt
1989 1052.853
1992 1013.949
1995 1735.111
1998 1959.001
2001 2311.100
2004 3224.881
2007 4793.357
2010 6880.418
2013 7281.225
2016 10942.393

oooo that is…. minimalistic. It appears when I run the chunk and only takes up the necessary space. It is perhaps a little skinny though…

Option 3 DT

The DT package is an interface for the javascript DataTables library.

summary %>%
  DT::datatable()

OK that one seems nice. I don’t love the default font (huh the font that displays in the Rmd is different to how it renders… why is that?). Its not too big (or too skinny) though and the search function could be useful if you were dealing with big tables. The decimal places are a bit of a problem though.

Challenges/Successes

Next steps