In jr-packages/jrTidyverse2: Jumping Rivers: Next Steps in the Tidyverse

First we'll load the data and the relevant packages

library(tidyverse)
data(okcupid, package = "jrTidyverse2")

This is a data set of dating profiles from the dating website OK Cupid. We're going to specifically focus on how drinking affects income in the land of OK Cupid.

Question 1

a) The column drinks corresponds to a persons answer about their drinking habits. What is the average income of each group of drinking habits? Save this as a data frame called drinks_in and have the column containing the average income called av_in Hint: use group_by() then summarise()

drinks_in = okcupid %>% 
  group_by(drinks) %>% 
  summarise(av_in = mean(income))

b) We can plot the average incomes using ggplot2

drinks_in %>% 
  ggplot(aes(x = drinks, y = av_in)) + 
  geom_point()

c) Previously we saw how to rename factors using fct_recode(). However, this will not work with missing values i.e. NA's. A function that will is fct_explicit_na(). Try running

x = c(1, 2, 3, NA)
(y = factor(x))

Notice how the NA isn't included in the factors?

fct_explicit_na(y, "unknown")

That will rename the NA factors as "Unknown". Use mutate() and fct_explicit_na() to rename the missing values to something more appropriate. Then plot the points using ggplot2.

drinks_in %>% 
  mutate(drinks = fct_explicit_na(drinks, "Unknown")) %>% 
  ggplot(aes(x = drinks, y = av_in)) + 
  geom_point()

d) Reorder the points from lowest average income to highest, then plot using ggplot2
Hint: use mutate() and fct_reorder()

drinks_in %>% 
  mutate(drinks = fct_explicit_na(drinks, "Unknown")) %>% 
  mutate(drinks = fct_reorder(drinks, av_in)) %>% 
  ggplot(aes(x = drinks, y = av_in)) + 
  geom_point()

e) Reorder the points order from people who drink least to people who drink most and then plot using ggplot2. Put the category "Unknown" where you deem appropriate.
Hint: use fct_relevel()

drinks_in %>% 
  mutate(drinks = fct_explicit_na(drinks, "Unknown")) %>% 
  mutate(drinks = fct_relevel(drinks, 
                              "Unknown", "not at all",
                              "rarely", "socially",
                              "often", "very often",
                              "desperately")) %>% 
  ggplot(aes(x = drinks, y = av_in)) + 
  geom_point()

f) Summarise the groups average income in the same way, but this time collapse "not at all" and "rarely" into "low", "socially" and "often" into "medium" and then "very often" and "desperately" into "high". Plot it using ggplot2.

okcupid %>% 
  mutate(drinks = fct_explicit_na(drinks, "Unknown")) %>% 
  mutate(drinks = fct_collapse(drinks,
                               Low = c("not at all", "rarely"),
                               Medium = c("socially", "often"),
                               High = c("very often", "desperately"))) %>%
  group_by(drinks) %>% 
  summarise(av_in = mean(income)) %>% 
  ggplot(aes(x = drinks, y = av_in)) + 
  geom_point()