Analysing my tips in cafés

tips-plots.R

library(tidyverse)

# Set Monday as the first day of the week.
options(lubridate.week.start = 1)

theme_set(
  theme_minimal() +
    theme(
      panel.grid.major.x = element_blank(),
      panel.grid.minor.x = element_blank(),
      panel.grid.minor.y = element_blank()
    )
)

tips <- read_csv("tips.csv") |>
  mutate(
    ratio = tip / bill,
    date  = dmy(date),
    wday  = wday(date, label = TRUE, abbr = TRUE),
    week  = week(date),
    year  = year(date)
  )

scale_y_percent <- function() {
  scale_y_continuous(labels = scales::label_percent(10), limits = c(0, .4))
}

# First, an elementary histogram of tip values in percent of bill total.
ggplot(tips, aes(ratio)) +
  geom_histogram(fill = "steelblue", binwidth = .05, boundary = 0, colour = "white") +
  scale_x_continuous(labels = scales::label_percent(10))

# Scatter plot of bill total vs. tip %.
ggplot(tips, aes(bill, ratio)) +
  geom_point(colour = "dodgerblue4", shape = 16, alpha = .5) +
  scale_y_percent()

# Time series of number of recorded tips per week.
ggplot(tips, aes(week)) +
  geom_bar(fill = "steelblue", width = 1) +
  facet_grid(rows = vars(year))

# Relative values of all tips each week as points.
ggplot(tips, aes(week, ratio)) +
  geom_point(colour = "dodgerblue4", shape = 16, alpha = .5) +
  geom_hline(yintercept = 0) +
  scale_y_percent() +
  facet_grid(rows = vars(year))

# Violin plot of relative values by day of week with overlaid jittered observations.
ggplot(tips, aes(wday, ratio)) +
  geom_violin(colour = "grey", draw_quantiles = .5) +
  geom_jitter(colour = "dodgerblue4", shape = 16, alpha = .5, width = .2) +
  scale_y_percent()