Use geom_label for fixed co-ords (instead of annotate) and review text (#72)

Author: alice-hannah

vignettes/articles/cookbook/_annotations.Rmd

@@ -2,6 +2,8 @@
 
 Labelling your chart is often preferable to using a legend, as often this relies on a user matching the legend to the data using colour alone. The legend can be removed from a chart by setting `legend = "none"` in `theme_sg()`.
 
+Either `geom_label()` or `geom_text()` can be used to annotate a chart. The main difference is that `geom_label()` draws a rectangle behind the text (white by default) and a border the same colour as the text (`label.size = NA` can be used to remove the border). `geom_text()` does not include a background and so text can be harder to read if it overlaps with other chart elements.
+
 The easiest way to add an annotation is to manually define the co-ordinates of the required position.
 
 ```{r annotations-data}
@@ -10,24 +12,26 @@ ann_data <- gapminder |>
 ```
 
 ```{r annotations-1}
-#| fig.alt = "A multiple line chart with colour co-ordinated line labels using sgplot theme and main palette."
+#| fig.alt = "A multiple line chart  using sgplot theme and main2 palette with lines labelled."
 
 ann_data |>
-  ggplot(aes(x = year, y = lifeExp, colour = country)) +
-  geom_line(linewidth = 1) +
+  ggplot() +
+  geom_line(aes(x = year, y = lifeExp, colour = country), linewidth = 1) +
   theme_sg(legend = "none") +
   scale_colour_discrete_sg() +
   scale_y_continuous(limits = c(0, 82),
                      breaks = seq(0, 80, 20),
                      expand = c(0, 0)) +
   scale_x_continuous(limits = c(1952, 2017),
                      breaks = seq(1952, 2007, 5)) +
-  annotate(geom = "label", x = 2008, y = 73, label = "China",
-           label.size = NA,
-           hjust = 0, vjust = 0.5) +
-  annotate(geom = "label", x = 2008, y = 79.4, label = "United Kingdom",
-           label.size = NA,
-           hjust = 0, vjust = 0.5) +
+  geom_label(x = 2008, y = 73, label = "China",
+             hjust = 0, 
+             vjust = 0.5,
+             label.size = NA) +
+  geom_label(x = 2008, y = 79.4, label = "United Kingdom",
+             hjust = 0, 
+             vjust = 0.5,
+             label.size = NA) +
   labs(
     x = "Year",
     y = NULL,
@@ -39,17 +43,13 @@ ann_data |>
 
 However, this makes the code difficult to reuse as values are hard coded and not automatically generated from the data. Automating the position of annotations is possible, but more fiddly. 
 
-The following examples use `geom_label()` to use values from the data to position annotations. `geom_label()` draws a rectangle behind the text (white by default) and a border the same colour as the text (`label_size = NA` can be used to remove the border). `geom_text()` is also an option for annotations, but this does not include a background and so can be harder for text to read if it overlaps with other chart elements. These functions also have `nudge` arguments that can be used to displace text to improve the positioning.
-
-Note that in the previous examples, `annotate()` also requires a geom (`label` or `text`). These operate in the same way as `geom_label()` and `geom_text()`, but as discussed, `annotate()` is only able to deal with fixed values.
-
+One way to automate this is to create a supplementary data frame with desired co-ordinates of the labels (see `ann_labs` in the example below). `nudge` arguments can be used to displace text to improve the positioning.
 
 ```{r annotations-2}
-#| fig.alt = "A multiple line chart with colour co-ordinated line labels using sgplot theme and main2 palette."
+#| fig.alt = "A multiple line chart  using sgplot theme and main2 palette with lines labelled."
 
 ann_labs <- ann_data |>
   group_by(country) |>
-  mutate(min_year = min(year)) |>
   filter(year == max(year)) |>
   ungroup()
 
@@ -102,5 +102,3 @@ ggplot(bar_data, aes(x = reorder(country, -pop), y = pop)) +
     caption = "Source: Gapminder"
   )
 ```
-
-**Note**: The `annotate()` function should be used to add annotations with manually defined positioning co-ordinates, whereas `geom_label()` and `geom_text()` should be used when using co-ordinates defined in a data frame. Although the reverse may work, text can appear blurry.