How to create area plots

This section will demo how to create a stacked area plot in fedplot style, using Figure 4.1 of the November 2022 FSR as a reference.

Example area plot

First, we load ggplot2, fedplot (which contains the sample dataset FSR_4_1), and scales.

#devtools::load_all()
library(ggplot2)
library(fedplot)
#> Warning in load_fed_font(): Cannot load font 'ITCFranklinGothic LT BookCn'; not
#> installed
library(scales)
library(forcats)

packageVersion("fedplot")
#> [1] '0.9.0'
head(FSR_4_1)
#> # A tibble: 6 × 3
#>   date       type                        value
#>   <date>     <chr>                       <dbl>
#> 1 2002-01-01 Other                        73.8
#> 2 2002-01-01 Securities lending           68.5
#> 3 2002-01-01 Commercial paper             64.8
#> 4 2002-01-01 Domestic money market funds  51.8
#> 5 2002-01-01 Repurchase agreements        30.2
#> 6 2002-01-01 Uninsured deposits           11.2

We can construct the area plot using standard ggplot2 functions:

FSR_4_1 |>
  ggplot(aes(x = date, y = value, fill=type)) +
  geom_area() +
  labs(y="Percent of GDP")

Now we customize it:

linewidth <- getOption("fedplot.linewidth_adj") * 0.25
font_size <- unit(getOption("fedplot.font_size") * 7L / 8L / ggplot2::.pt, "bigpts")
font_family <- getOption("fedplot.font_family")
maxdate <- max(FSR_4_1$date)

FSR_4_1 |>
  dplyr::mutate(
    type = dplyr::case_when(
      type == "Other" ~ "1. Other",
      type == "Securities lending" ~ "2. Securities lending",
      type == "Commercial paper" ~ "3. Commercial paper",
      type == "Domestic money market funds" ~ "4. Domestic money market funds",
      type == "Repurchase agreements" ~ "5. Repurchase agreements",
      type == "Uninsured deposits" ~ "6. Uninsured deposits",
      .default = "N/A")) |>
  dplyr::mutate(type=forcats::fct_rev(factor(type))) |> 
  #dplyr::mutate(type=factor(type)) |> 
  ggplot(aes(x = date, y = value, fill=type)) +
  geom_area(color="black", linewidth=linewidth, key_glyph=draw_key_square, outline.type="full") + # TBH looks better with color="white"...
  labs(y="Billions of dollars") +
  geom_hline_zero() +
  scale_x_date(minor_breaks=seq(from=as.Date("2002-01-01"), to=as.Date("2023-01-01"), by="1 years"),
               breaks=seq(from=as.Date("2002-06-30"), to=as.Date("2022-06-30"), by="4 years"),
               date_labels="%Y",
               expand=expansion(mult=.05)) +
  scale_y_continuous(sec.axis = dup_axis(),
                   breaks = seq(0, 500, by=100),
                   limits = c(0, 500),
                   expand = expansion(mult=0),
                   labels = scales::label_number(style_negative = "minus")) +
  annotate_last_date(nudge_y = 350) +
  
  # Demo of how we could start implementing the number annotations...
  # See also: https://stackoverflow.com/questions/10393956/add-direct-labels-to-ggplot2-geom-area-chart
  geom_segment(aes(x=maxdate, y=330, xend=maxdate+120, yend=360), linewidth=linewidth, lineend="round") +
  annotate("text", x=maxdate+150, y=360, label="1", size=font_size, family=font_family) +
  geom_segment(aes(x=maxdate, y=300, xend=maxdate+120, yend=300), linewidth=linewidth, lineend="round") +
  annotate("text", x=maxdate+150, y=300, label="2", size=font_size, family=font_family) +
  
  # Hack to add an outline around the stacked area geom
  #geom_line(aes(ymax=value), position="stack", linewidth=linewidth) +
  
  guides(fill=guide_legend(ncol=2, reverse=TRUE)) +
  theme_fed(legend_position = c(.38, 1.1),
            fill_palette=fedplot::bsvr_colors,
            size='wide')

Note that to achieve the exact ordering used in the FSR (where the stacked areas have a reverse order but the legend has an alphabetical order) we can use several two tricks:

1. Use forcats::fct_rev to reverse the order of the type variable (see also this link).
2. Call guide_legend() with the reverse=TRUE option.

Lastly, we want to export the chart so it matches the required image characteristics:

save_plot('areaplot', extension='all')
#> saved 'areaplot.pdf' (1.69296638166501x6.13546463673551; dpi=300)
#> saved 'areaplot.eps' (1.69296638166501x6.13546463673551; dpi=300)
#> saved 'areaplot.png' (1.69296638166501x6.13546463673551; dpi=600)

After exporting through save_plot, the chart looks like this:

Pending tasks:

1. Add the pattern area, using ggpattern.
2. Get data in percent of GDP instead of USD. Also check the actual values for the OTHER category.
3. Get correct colors
4. Add numbers to legends and annotations
5. Add black stripped area for the period from 2008:Q4 to 2012:Q4