Update and expand pub quality figures

Include using functions to make the same graph repeatedly

Author: ethanwhite

materials/publication-figures-R.md

@@ -4,8 +4,8 @@ element: notes
 title: Publication quality figures
 language: R
 ---
- 
-> * Have students install `devtools` and `patchwork` (using `devtools`)
+
+> * Have students install `ggplot2`, `readr`, `patchwork`
 > * Open https://ggplot2.tidyverse.org/reference/ggtheme.html in browser
 
 ### File formats
@@ -18,16 +18,17 @@ language: R
   * PNG is a good compromise format
 * Vector
   * Right choice for plots, line drawings
+  * Describes the objects that make up the image, including their shapes, colors, and positions
   * Provides infinite scaling
-    * EPS, AI, PDF, SVG
+    * EPS (printers), AI, PDF, SVG (web)
 
 * Save in different file formats using different extensions
 
 ```r
 library(ggplot2)
 library(readr)
 
-trees <- read_tsv("https://ndownloader.figshare.com/files/5629536")
+trees <- read_tsv("https://ndownloader.figshare.com/files/5629536", na = c("", "NA", "missing", "MISSING", "?", "3.3."))
 
 ggplot(trees, aes(x = HEIGHT, y = CIRC)) +
   geom_point()
@@ -58,13 +59,16 @@ ggsave("acacia_size_scaling.png", dpi = 300)
 ggsave("acacia_size_scaling.png", dpi = 30)
 ```
 
+* When submitting to a journal first think about how large the image will be
+* Then set the dpi to at least the journal's minimum resolution
+
 ### Color palettes
 
 * Choose colors that work well for everyone
-  * People with different kinds of color blindness
+  * People with different kinds of color vision
   * People who printed your paper out without a color printer
 * Need to be correctly interpreted
-* Viridis is a new color scale that is designed to provide a good set of default
+* Viridis is a new(ish) color scale that is designed to provide a good set of default
   colors for addressing all of these concerns
 
 ```r
@@ -88,7 +92,7 @@ ggplot(trees, aes(x = HEIGHT, y = CIRC, color = HEIGHT)) +
 ### Themes
 
 * Can customize every aspect of plots in `ggplot`
-* Themes are an easy way to change the overall look of figures 
+* Themes are an easy way to change the overall look of figures
 * These can be used to make coordinated changes to groups of options
 
 ```r
@@ -117,17 +121,6 @@ ggsave("species_scaling.jpg", species_scaling)
 
 * Often want to combine multiple distinct plots into a single figure
 * Two popular packages for this, `patchwork` and `cowplot`
-
-* `patchwork` is not on CRAN so install using `devtools`
-* `devtools` lets us install packages from a variety of sources, including
-  GitHub, one of the major hubs of software development
-
-```r
-install.packages('devtools')
-library(devtools)
-install_github('thomasp85/patchwork')
-```
-
 * `patchwork` works by "adding" plots to one another
 
 ```r
@@ -171,4 +164,79 @@ height_dist <- height_dist +
   theme(legend.position='none')
 
 height_dist + species_scaling + plot_layout(ncol = 1, heights = c(1, 5))
-```
+```
+
+### Reusing custom plots
+
+* Once you've created a plot you might want to reuse the same basic plotting code
+* The most common way folks do this is to copy and paste the code and change the pieces they want to change
+* A better way is using functions
+* Let's start with just the first piece of our graph
+
+```r
+make_plot <- function(data, xcolumn, ycolumn, colorcolumn) {
+  species_scaling <- ggplot(data, aes(x = xcolumn, y = ycolumn, color = colorcolumn)) +
+    geom_point() +
+    scale_color_viridis_d()
+}
+
+plot_circ <- make_plot(trees, HEIGHT, CIRC, SPECIES)
+plot_circ
+```
+
+* This doesn't work, but why?
+* There is an extra step we need to take when working with tidyverse functions that work with "data variables", i.e., names of columns that are not in quotes
+* These functions use tidy evaluation, a special type of non-standard evaluation
+* This basically means they do fancy things under the surface to make them easier to work with
+* But it means they don't work if we just pass things to functions in the most natural way
+* To fix this we have to tell our code which inputs/arguments are this special type of data variable
+* We do this by "embracing" them in double braces
+
+```r
+make_plot <- function(data, xcolumn, ycolumn, colorcolumn) {
+  {% raw %}scaling <- ggplot(data, aes(x = {{ xcolumn }}, y = {{ ycolumn }}, color = {{ colorcolumn }})) +{% endraw %}
+    geom_point() +
+    scale_color_viridis_d()
+}
+
+plot_circ <- make_plot(trees, HEIGHT, CIRC, SPECIES)
+plot_circ
+```
+
+* Now let's add the rest of our plot
+
+```r
+make_plot <- function(data, xcolumn, ycolumn, colorcolumn) {
+{% raw %}scaling <- ggplot(data, aes(x = {{ xcolumn }}, y = {{ ycolumn }}, color = {{ colorcolumn }})) +{% endraw %}
+  geom_point() +
+  scale_color_viridis_d()
+
+{% raw %}distribution <- ggplot(data, aes(x = {{ xcolumn }}, fill = {{ colorcolumn }})) +{% endraw %}
+    geom_histogram() +
+    scale_fill_viridis_d() +
+    theme_void() +
+    theme(legend.position='none')
+
+  distribution + scaling + plot_layout(ncol = 1, heights = c(1, 5))
+}
+
+plot_circ <- make_plot(trees, HEIGHT, CIRC, SPECIES)
+plot_circ
+```
+
+* Having done this we can now plot whatever we want
+
+```r
+plot_axis1 <- make_plot(trees, HEIGHT, AXIS_1, SPECIES)
+plot_axis1
+```
+
+```r
+plot_axis1_year <- make_plot(trees, HEIGHT, AXIS_1, as.factor(YEAR))
+plot_axis1_year
+```
+
+```r
+plot_axis1_axis2_year <- make_plot(trees, AXIS_1, AXIS_2, as.factor(YEAR))
+plot_axis1_axis2_year
+```