In Athanasiamo/swc.tidyverse: Introduction to Tidyverse

library(learnr)
library(gradethis)

knitr::opts_chunk$set(
  echo = FALSE,
  exercise.warn_invisible = FALSE
)

# enable code checking
tutorial_options(
  exercise.checker = grade_learnr,
  exercise.lines = 20,
  exercise.reveal_solution = TRUE
)

Challenge 1

1a

In our new summary function, alter the code so that the columns are no longer prefixed with value_.

penguins %>% 
  pivot_longer(ends_with("mm")) %>% 
  group_by(name) %>% 
    summarise(across(value, 
                   .fns = list(mean = mean,
                               sd = sd,
                               min = min,
                               max = max), 
                   na.rm = TRUE,
            n = length(species)
  ) 

penguins %>% 
  pivot_longer(ends_with("mm")) %>% 
  group_by(name) %>% 
    summarise(across(value, 
                   .fns = list(mean = mean,
                               sd = sd,
                               min = min,
                               max = max), 
                   na.rm = TRUE,
                   .names = "{.fn}"),
            n = length(species)
  ) 

grade_code(
  correct = random_praise(),
  incorrect = random_encouragement()
)

Try using the `.names` argument in across.

The internal placeholder for the function names in across is `.fn`.

1b

Adapt the code so that NAs in the values are removed during the pivot_longer, and try to remove na.rm = TRUE from the across. Does that work? Why?

penguins %>% 
  pivot_longer(ends_with("mm")) %>% 
  group_by(name) %>% 
    summarise(across(value, 
                   .fns = list(mean = mean,
                               sd = sd,
                               min = min,
                               max = max), 
                   .names = "{.fn}"),
            n = length(species)
  ) 

penguins %>% 
  pivot_longer(ends_with("mm"),
               values_drop_na = TRUE) %>% 
  group_by(name) %>% 
    summarise(across(value, 
                   .fns = list(mean = mean,
                               sd = sd,
                               min = min,
                               max = max), 
                   .names = "{.fn}"),
            n = length(species)
  ) 

grade_code(
  correct = random_praise(),
  incorrect = random_encouragement()
)

Check out the `values_drop_na` argument in pivot longer

1c

Adapt the code so that the "n" is captured within the across function list, like the other four functions. How is the output different?

penguins %>% 
  pivot_longer(ends_with("mm")) %>% 
  group_by(name) %>% 
    summarise(across(value, 
                   .fns = list(mean = mean,
                               sd = sd,
                               min = min,
                               max = max), 
                   .names = "{.fn}"),
            n = length(species)
  ) 

penguins %>% 
  pivot_longer(ends_with("mm"),
               values_drop_na = TRUE) %>% 
  group_by(name) %>% 
    summarise(across(value, 
                   .fns = list(mean = mean,
                               sd = sd,
                               min = min,
                               max = max,
                               n = length), 
                   .names = "{.fn}")
  )

grade_code(
  correct = random_praise(),
  incorrect = random_encouragement()
)

1d

Try grouping by more variables. Is the outcome as you expect?

penguins %>% 
  pivot_longer(ends_with("mm"),
               values_drop_na = TRUE) %>% 
  group_by(name) %>% 
    summarise(across(value, 
                   .fns = list(mean = mean,
                               sd = sd,
                               min = min,
                               max = max,
                               n = length), 
                   .names = "{.fn}")
  )

Challenge 2

2a

Create a bar chart based om the penguins summary data, where the mean values are on the x axis and species are on the y axis. Make sure to dodge the bar for easier comparisons. Create subplots on the different metrics.

__ %>% 
  ggplot(aes(x = __, 
             y = __,
             fill = __)) +
  geom_bar(stat = __) +
  facet___(~ __, scales = "free")

penguins_sum %>% 
  ggplot(aes(x = mean, 
             y = species,
             fill = island)) +
  geom_bar(stat = "identity",
           position = "dodge") +
  facet_wrap(~ name, scales = "free")

grade_code(
  correct = random_praise(),
  incorrect = random_encouragement()
)

2b

Adapt the code the plot the standard deviation rather than the mean

penguins_sum %>% 
  ggplot(aes(x = mean, 
             y = species,
             fill = island)) +
  geom_bar(stat = "identity",
           position = "dodge") +
  facet_wrap(~ name, scales = "free")

penguins_sum %>% 
  ggplot(aes(x = sd, 
             y = species,
             fill = island)) +
  geom_bar(stat = "identity",
           position = "dodge") +
  facet_wrap(~ name, scales = "free")

grade_code(
  correct = random_praise(),
  incorrect = random_encouragement()
)

2c

Swap around the plot to explore different ways of looking at the same data. Try swapping the axes, so that x is on the y, and vice versa. Or Try having species on one axis and using island as colour fill. What variation do you think best shows the differences between groups?

penguins_sum %>% 
  ggplot(aes(x = species, 
             y = sd,
             fill = island)) +
  geom_bar(stat = "identity",
           position = "dodge") +
  facet_wrap(~ name, scales = "free")

2d

Try adding an error bar that indicates the standard deviation of the measurement, while the bars indicate the mean values. What does the geom_errorbar need to plot the error bars?

penguins_sum %>% 
  ggplot(aes(x = mean, 
             y = species,
             fill = island,
             group = __)) +
  geom_bar(stat = "identity",
           position = "dodge") +
  geom___(aes(__ = _ ,
              __ = _),
          position = "dodge") +
  facet_wrap(~ name, scales = "free")

penguins_sum %>% 
  ggplot(aes(x = mean, 
             y = species,
             fill = island,
             group = island)) +
  geom_bar(stat = "identity",
           position = "dodge") +
  geom_errorbar(aes(xmin = mean - sd ,
                    xmax = mean + sd),
                position = "dodge") +
  facet_wrap(~ name, scales = "free")

grade_code(
  correct = random_praise(),
  incorrect = random_encouragement()
)

the geom for error bars is called `geom_errorbar`

Error bars are lines between two points. Try looking for a way to ass the minimum and maximum values for the x-axis.

2e

Try pivoting the data even longer! Pivot longer all the stats columns so that the column names are in a column named "stat". Now, create another plot using the "value" column on the y-axis, and creating subplots based on both the observation name AND the stat!

penguins_sum %>% 
  pivot_longer(all_of(c(__)),
               __ = __) %>% 
  ggplot(aes(x = mean, 
             y = species,
             fill = island,
             group = __)) +
  geom_bar(stat = "identity",
           position = "dodge") +
  facet___( __ ~ name, scales = "free")

penguins_sum %>% 
  pivot_longer(all_of(c("mean", "sd", "min", "max")),
               names_to = "stat") %>% 
  ggplot(aes(x = species, 
             y = value,
             fill = island)) +
  geom_bar(stat = "identity",
           position = "dodge") +
  facet_grid(stat ~ name, scales = "free")

grade_code(
  correct = random_praise(),
  incorrect = random_encouragement()
)

Try adding the names of the columns to pivot in quotation marks within the c().

Try the facet_grid function, which takes the syntax `facet_grid(row_column ~ columns_column).

Athanasiamo/swc.tidyverse documentation built on Dec. 17, 2021, 9:48 a.m.