Toronto Shelters (Tidy Tuesday 2020-12-01)

###################################################################################################
# Toronto Shelters - Average nightly number of people served in Toronto's shelter system.
# Created by: reddit.com/user/brianhaas19

# Link to data
# https://github.com/rfordatascience/tidytuesday/blob/master/data/2020/2020-12-01

# Setup
library(tidyverse)
library(scales)
library(patchwork)
library(ggrepel)
library(ggsci)
theme_set(theme_bw())

# Load data
shelters <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2020/2020-12-01/shelters.csv')

# Clean data
shelters <- shelters %>%
  rename(date = occupancy_date)

# Visualize
### Recommended parameters for R notebook: fig.height=7, fig.width=9
line_size = 0.5
p1 <- shelters %>%
  filter(is.finite(capacity)) %>% # to be consistent with the second plot
  mutate(month = lubridate::floor_date(date, "month")) %>%
  group_by(month, sector) %>%
  mutate(num_days = n_distinct(date)) %>% # number of days in the month for calculating nightly average
  summarise(occupancy = sum(occupancy)/num_days) %>%
  distinct() %>%
  ungroup() %>% # ungroup() here so that the next line to create the labels works
  mutate(label = if_else(month == max(month), as.character(sector), NA_character_)) %>%
  ggplot(aes(month, occupancy, group = sector, color = sector)) +
  geom_line(size = line_size, show.legend = FALSE) +
  geom_label_repel(aes(label = label), seed = 1, show.legend = FALSE, size = 4, label.padding = 0.15,
                   nudge_x = 0, na.rm = TRUE) + # *1
  scale_y_continuous(labels = comma) +
  ggsci::scale_color_jama() +
  labs(x = NULL, y = "Number of People", title = expression(paste("Average nightly ", bold(occupancy), " by ", bold(sector), ":"))) +
  theme(plot.title = element_text(size = 12))

p2 <- shelters %>%
  filter(is.finite(capacity)) %>%
  mutate(month = lubridate::floor_date(date, "month")) %>%
  group_by(month) %>%
  mutate(num_days = n_distinct(date)) %>%
  summarise(occupancy = sum(occupancy)/num_days,
            capacity = sum(capacity)/num_days) %>%
  distinct() %>%
  pivot_longer(cols = c(occupancy, capacity), names_to = "type") %>%
  ungroup() %>% # ungroup() here so that the next line to create the labels works
  mutate(label = if_else(month == max(month), if_else(type == "occupancy", "Occupancy", "Capacity"), NA_character_)) %>%
  ggplot(aes(month, value, color = type)) +
  geom_line(size = line_size, show.legend = FALSE) +
  geom_label_repel(aes(label = label), seed = 1, show.legend = FALSE, size = 4, label.padding = 0.15,
                   nudge_x = 0, na.rm = TRUE) + # *1
  scale_y_continuous(labels = comma) +
  scale_color_manual(values = c("#80796B", "#6A6599")) + # these are the remaining two colors from ggsci::scale_color_jama() not used above
  labs(x = NULL, y = "Number of People", title = expression(paste("Average nightly ", bold(occupancy), " and ", bold(capacity), " - all sectors:"))) +
  theme(plot.title = element_text(size = 12))

# Use patchwork package to combine the plots:
P <- p1/p2 + plot_annotation(title = expression(paste(bold("Toronto Shelters"), " - Average nightly number of people served in Toronto's shelter system.")),
                        caption = expression(paste("Created by: ", italic("reddit.com/user/brianhaas19"), "\tData source: ", italic("https://github.com/rfordatascience/tidytuesday/blob/master/data/2020/2020-12-01"))))
P

# Uncomment this line to save to disk:
# ggsave(str_c(getwd(), "/toronto_shelters.png"),
#        plot = P, height = 7, width = 9)
###################################################################################################