conneced_scatterplot.qmd

conneced_scatterplot.qmd

```{r}
setwd(here::here('connected_scatterplots'))

library(tidyverse)
library(lubridate)
library(patchwork)
library(showtext)

showtext_auto()
showtext_opts(dpi = 300)
font_add_google('Fira Sans', 'firasans')

camcorder::gg_record(
  dir = 'img',
  width = 16,
  height = 9,
  units = 'cm'
)

# FiveThirtyEight Datat
# https://projects.fivethirtyeight.com/biden-approval-rating/
joe_approvals <- read_csv('approval_topline.csv') |> 
  filter(subgroup == 'All polls', president == 'Joe Biden') |> 
  select(
    date = modeldate, 
    approval = approve_estimate, 
    disapproval = disapprove_estimate
  ) |> 
  mutate(date = parse_date(date, format = "%m/%d/%Y")) |> 
  filter(
    between(
      date,
      lubridate::make_date(2022, 1, 3),
      lubridate::make_date(2022, 10, 17)
    )
  )
  


# Federal Reserve Data
# https://fred.stlouisfed.org/series/GASREGW
gas_prices <- read_csv('GASREGW.csv') |> 
  janitor::clean_names() |> 
  filter(
    between(
      date,
      lubridate::make_date(2022, 1, 3),
      lubridate::make_date(2022, 10, 17)
    )
  ) |> 
  mutate(
    gasregw = parse_number(gasregw)
  )
```


# Approval plots

```{r}
weekly_joe_approvals <- joe_approvals  |> 
  filter(date %in% (make_date(2022, 1, 03) + days(7 * 0:100)))

special_date <- lubridate::make_date(2022, 06, 20)

duplicated_row <- weekly_joe_approvals |> 
  filter(date >= special_date) |> 
  arrange(date) |> 
  slice(n = 1) |> 
  mutate(after_special_date = FALSE)

color_palette <- c(thematic::okabe_ito(1), 'black')

joe_with_duplicate <- weekly_joe_approvals |> 
  mutate(after_special_date = (date >= special_date)) |> 
  bind_rows(duplicated_row)

approval_plot <- joe_with_duplicate  |> 
  ggplot(aes(
    date, 
    approval,
    color = after_special_date 
  )) +
  geom_line(linewidth = 0.75) +
  theme_minimal() +
  theme(
    legend.position = 'none',
    plot.title.position = 'plot',
    panel.grid.minor = element_blank()
  ) +
  coord_cartesian(ylim = c(35, 45)) +
  scale_color_manual(values = color_palette) +
  scale_x_date(labels = scales::label_date(format = "%b %d", locale = 'en_US')) +
  labs(x = element_blank(), y = element_blank(), title = 'Biden Approval Rating')
```



# Gas plot


```{r}
duplicated_row_gas <- gas_prices |> 
  filter(date >= special_date) |> 
  slice(n = 1) |> 
  mutate(after_special_date = FALSE)

gas_with_duplicate <- gas_prices |> 
  mutate(after_special_date = (date >= special_date)) |> 
  bind_rows(duplicated_row_gas)

gas_plot <- gas_with_duplicate |> 
  ggplot(aes(
    date, 
    gasregw,
    color = after_special_date 
  )) +
  geom_line(linewidth = 0.75) +
  theme_minimal() +
  theme(
    legend.position = 'none',
    plot.title.position = 'plot',
    panel.grid.minor = element_blank()
  ) +
  coord_cartesian(ylim = c(3, 5)) +
  scale_y_continuous(labels = scales::label_dollar()) +
  scale_x_date(labels = scales::label_date(format = "%b %d", locale = 'en_US')) +
  scale_color_manual(values = color_palette) +
  labs(x = element_blank(), y = element_blank(), title = 'Gas prices (Price per Gallon)')
```


# Connected Scatter Plot


```{r}
date_labels <- c(
  make_date(2022, 1, 3),
  make_date(2022, 3, 14),
  make_date(2022, 6, 13),
  make_date(2022, 7, 18),
  make_date(2022, 10, 17)
)

joined_data <- joe_with_duplicate |> 
  left_join(gas_with_duplicate) |> 
  arrange(date) |> 
  mutate(
    label = if_else(
      date %in% date_labels, 
      scales::label_date(format = "%b %d", locale = 'en_US')(date),
      ''
    )
  )

connected_plot <- joined_data |> 
  ggplot(aes(approval, gasregw)) +
  geom_path(aes(col = after_special_date), linewidth = 0.75) +
  ggrepel::geom_text_repel(
    aes(label = label), 
    min.segment.length = 0.25, 
    box.padding = 1, 
    max.overlaps = 100,
    color = 'grey30',
    size = 2.5,
    family = 'firasans'
  ) +
  geom_point(aes(fill = after_special_date), col = 'black', shape = 21, size = 1.5) +
  theme_minimal(base_size = 10) +
  theme(
    legend.position = 'none',
    plot.title.position = 'plot',
    panel.grid.minor = element_blank()
  ) +
  coord_cartesian(ylim = c(2.75, 5.25), xlim = c(35, 45)) +
  scale_y_continuous(labels = scales::label_dollar()) +
  scale_color_manual(values = color_palette) +
  scale_fill_manual(values = color_palette) +
  labs(
    x = 'Approval Rating', 
    y = 'Price per gallon', 
    title = 'Concurrent Relationship between Gas Prices\nand Biden Approval Rating',
    subtitle = 'Each point a week in 2022'
  )
connected_plot
```



# Assemble plots

```{r}
(((gas_plot / approval_plot) | connected_plot) &
  theme(text = element_text(color =  'grey30', size = 8, family = 'firasans')) ) +
  plot_annotation(
    title = 'Gas prices vs. Biden approval',
    subtitle = 'From the beginning of the year to mid-October',
    caption = 'Sources: WP article by Philip Bump, FiveThirtyEight, Federal Reserve Bank of St. Louis\nConnected scatterplot design: @jschwabish\nggplot2 remake: @rappa753',
    theme = theme(
      text = element_text(color =  'grey30', size = 12, family = 'firasans'),
      plot.caption = element_text(size = 6, lineheight = 1)
    )
  ) 
```