articles

R/add-general.R

@@ -236,15 +236,15 @@ ff_value <- function(.fun, .count = FALSE) {
     if (.count) {
       plot <- plot +
         ggplot2::stat_count(ggplot2::aes(label = format_number(ggplot2::after_stat(count), accuracy = accuracy, scale_cut = scale_cut)),
-                            geom = "text", vjust = vjust, hjust = hjust, size = size, position = position, ...)
+                            geom = "text", vjust = vjust, hjust = hjust, size = size, position = position, show.legend = FALSE, ...)
     } else {
       if ((stringr::str_sub(ptype, 2, 2) == "c")) {
         plot <- plot + ggplot2::stat_summary(ggplot2::aes(label = format_number(ggplot2::after_stat(y), accuracy = accuracy, scale_cut = scale_cut)),
-                              fun = .fun, geom = "text", vjust = vjust, hjust = hjust, size = size, position = position, ...)
+                              fun = .fun, geom = "text", vjust = vjust, hjust = hjust, size = size, position = position, show.legend = FALSE, ...)
       }
       if ((stringr::str_sub(ptype, 1, 1) == "c")) {
           plot <- plot + ggplot2::stat_summary(ggplot2::aes(label = format_number(ggplot2::after_stat(x), accuracy = accuracy, scale_cut = scale_cut)),
-                              fun = .fun, geom = "text", vjust = vjust, hjust = hjust, size = size, position = position, ...)
+                              fun = .fun, geom = "text", vjust = vjust, hjust = hjust, size = size, position = position, show.legend = FALSE, ...)
       }
     }
     plot

_pkgdown.yml

@@ -15,9 +15,9 @@ articles:
 - title: Articles
   navbar: ~
   contents:
+  - articles/Advanced-plotting
   - articles/Visualizing-data
   - articles/Color-schemes
-  - articles/Advanced-plotting
   - articles/Design-principles
 
 reference:
@@ -32,10 +32,10 @@ reference:
 - subtitle: Data points & amounts
 - contents:
   - starts_with("add_data_points")
-  - starts_with("add_line")
-  - starts_with("add_heatmap")
   - starts_with("add_count")
   - starts_with("add_sum")
+  - starts_with("add_heatmap")
+  - starts_with("add_line")
 - subtitle: Central tendency
 - contents:
   - starts_with("add_mean")

tests/testthat/test-add-misc.R

@@ -14,6 +14,7 @@ test_that("add annotation", {
     vdiffr::expect_doppelganger("add title and caption", .)
 })
 
+# violins failing in remote check
 # test_that("add boxplot and violin works", {
 #   study %>%
 #     tidyplot(x = treatment, y = score, color = treatment) %>%

tests/testthat/test-adjust.R

@@ -106,6 +106,7 @@ test_that("plotmath expressions work", {
     vdiffr::expect_doppelganger("plotmath expression x axis labels", .)
 })
 
+# failing in remote check
 # test_that("adjust plot area size work", {
 #   study %>%
 #     tidyplot(x = dose, y = score, color = group) %>%

vignettes/articles/Advanced-plotting.Rmd

@@ -271,3 +271,5 @@ time_course %>%
   add_title("orientation = y")
 ```
 
+# Padding
+

vignettes/articles/Visualizing-data.Rmd

@@ -20,12 +20,209 @@ knitr::opts_chunk$set(
 ```{r setup}
 library(tidyplots)
 ```
+
+# Data points
+
+Plotting the raw data points is probably the most bare bone way to visualize a dataset. the corresponding function in tidyplots is called `add_data_points()`.
+
+```{r}
+animals %>% 
+  tidyplot(x = weight, y = size) %>% 
+  add_data_points()
+```
+
+In the above example some data points appear to overlap other points. To account for this so called _overplotting_, you might want to add a thin white border around the points. This is achieved by setting the parameter `confetti = TRUE`.
+
+```{r}
+animals %>% 
+  tidyplot(x = weight, y = size) %>% 
+  add_data_points(confetti = TRUE)
+```
+
+Another way is to make the points transparent using the `alpha` parameter.
+
+```{r}
+animals %>% 
+  tidyplot(x = weight, y = size) %>% 
+  add_data_points(alpha = 0.3)
+```
+
+Or by changing the plotting symbol to an open `shape`.
+
+```{r}
+animals %>% 
+  tidyplot(x = weight, y = size) %>% 
+  add_data_points(shape = 1)
+```
+
+However, data points can also be used when plotting a _discrete variable_ like `treatment` against a _continuous variable_ like `score`.
+
+```{r}
+study %>% 
+  tidyplot(x = treatment, y = score) %>% 
+  add_data_points()
+```
+
+To avoid overplotting in this scenario, there are two additional options. You can add some random noise or _jitter_ to the y position.
+
+```{r}
+study %>% 
+  tidyplot(x = treatment, y = score) %>% 
+  add_data_points_jitter()
+```
+
+Alternatively, you can use an algorithm that keeps the points centered and just moves potentially overlapping points to the sides.
+
+```{r}
+study %>% 
+  tidyplot(x = treatment, y = score) %>% 
+  add_data_points_beeswarm()
+```
+
 # Amounts
 
-# Distributions
+For some datasets, it makes sense to `count` or `sum` up, data points in order to arrive to a conclusion. As one example, let's have a look at the `spendings` dataset.
+
+```{r results='markup'}
+spendings
+```
+
+As you can see, this dataset contains family spendings over a time period of 15 days in October. Here, it might be informative to see which spending categories are reoccurring and which are just one time spendings.
+
+```{r}
+spendings %>% 
+  tidyplot(x = category) %>% 
+  add_count_bar()
+```
+
+One thing to note here is that the x axis labels are overlapping and are thus unreadable. There are at least two possible solutions for this. One is to swap the x and y axis.
 
 ```{r}
+spendings %>% 
+  tidyplot(y = category) %>% 
+  add_count_bar()
+```
+
+The other one is to rotate the x axis labels.
+
+```{r}
+spendings %>% 
+  tidyplot(x = category) %>% 
+  add_count_bar() %>% 
+  adjust_x_axis(rotate_labels = TRUE)
+```
+
+Now we can appreciate that this family had reoccurring spendings for _Food_ but just one spending for _Housing_.
+
+Next, we ask the question how much was spend on each of the categories by plotting the `sum` amount.
+
+```{r}
+spendings %>% 
+  tidyplot(x = category, y = amount, color = category) %>% 
+  add_sum_bar() %>% 
+  adjust_x_axis(rotate_labels = TRUE)
+```
+
+Note that we had to introduce the parameter `y = amount` in the `tidyplot()` function to make it clear which variable should be summed up.
+
+I also added `color = category` in the `tidyplot()` function to have the variable `category` encoded by different colors.
+
+Since the data labels for the variable `category` are now duplicated in the plot, one could argue that it would be justified to remove the duplicated information on the x axis.
+
+```{r}
+spendings %>% 
+  tidyplot(x = category, y = amount, color = category) %>% 
+  add_sum_bar() %>% 
+  adjust_x_axis(rotate_labels = TRUE) %>% 
+  remove_x_axis_labels() %>% 
+  remove_x_axis_title() %>% 
+  remove_x_axis_ticks()
+```
 
+Note that besides the x axis labels, I also removed the x axis ticks and x axis title to achieve a clean look.
+
+Of course you are free to play around with different graphical representations of the sum values. Here is an example of a lollipop plot constructed from a thin `bar` and a `dot`.
+
+```{r}
+spendings %>% 
+  tidyplot(x = category, y = amount, color = category) %>% 
+  add_sum_bar(width = 0.03) %>% 
+  add_sum_dot() %>% 
+  add_sum_value(accuracy = 1) %>% 
+  adjust_x_axis(rotate_labels = TRUE) %>% 
+  remove_x_axis_labels() %>% 
+  remove_x_axis_title() %>% 
+  remove_x_axis_ticks()
+```
+
+I also added the sum value as text label using the `add_sum_value()` function. 
+
+# Heatmaps
+
+Heatmaps are a great way to plot a _continuous variable_ across to two _discrete variables_. To exemplify this, we will have a look at the `gene_expression` dataset.
+
+```{r results='markup'}
+gene_expression %>% 
+  dplyr::glimpse()
+```
+
+I used the `dplyr::glimpse()` function from the dplyr package get a nice overview of all variables including in the dataset. We will start by plotting the `expression` values of each `external_gene_name` across the `sample` variable.
+
+```{r}
+gene_expression %>% 
+  tidyplot(x = sample, y = external_gene_name, color = expression) %>% 
+  add_heatmap()
+```
+
+One thing to note here is that the y axis labeks are overlapping. So let's increase the height of the plot area from 50 to 80 mm.
+
+```{r fig.height=5}
+gene_expression %>% 
+  tidyplot(x = sample, y = external_gene_name, color = expression) %>% 
+  add_heatmap() %>%
+  adjust_plot_area_size(height = 80)
+```
+
+The next thing to note is that some of the rows like _Map1a_ and _Kif1a_ show very high expression while others show much lower values. Let's apply a classical technique to reserve the variations in the color for differences within each  row. This is done by calculating _row z scores_ for each row individually. Luckily, tidyplots does this for us when setting the parameter `scale = "row"` withing the `add_heatmap()` function call.
+
+```{r}
+gene_expression %>% 
+  tidyplot(x = sample, y = external_gene_name, color = expression) %>% 
+  add_heatmap(scale = "row") %>%
+  adjust_plot_area_size(height = 80)
+```
+
+Now it much easier to appreciate the dynamics of individual genes across the samples on the x axis.
+
+However, the rows appear to be mixed. Some having rather high expression in the "Eip" samples while others have high value in the "Hip" samples. Conveniently, in the dataset there is a variable called `direction`, which is either "up" or "down". Let's use this variable to sort our y axis.
+
+```{r}
+gene_expression %>% 
+  tidyplot(x = sample, y = external_gene_name, color = expression) %>% 
+  add_heatmap(scale = "row") %>%
+  adjust_plot_area_size(height = 80) %>% 
+  sort_y_axis_labels(direction)
+```
+
+This starts looking like a classical gene expression heatmap
+
+# Central tendency
+
+
+
+```{r}
+
+```
+
+# Dispersion & uncertainty
+
+```{r}
+
+```
+
+# Distribution
+
+```{r}
 distributions %>% 
   tidyplot(name, value) %>% 
   add_mean_bar(alpha = 0.3) %>% 
@@ -59,27 +256,11 @@ energy %>%
   tidyplot(x = energy_type, y = power, color = energy_type) %>% 
   add_violin() %>% 
   add_data_points_beeswarm(jitter_width = 0.8, alpha = 0.3)
-
 ```
 
-
-# Proportions
+# Proportion
 
 ```{r}
-
-# themes
-
-p4 <- 
-  energy %>% 
-  tidyplot(year, power, color = energy_source) %>% 
-  add_barstack_absolute()
-
-p4
-p4 %>% theme_minimal_y()
-p4 %>% theme_ggplot2()
-
-# proportions
-
 energy %>% 
   tidyplot(year, power, color = energy_source) %>% 
   add_barstack_absolute()
@@ -163,7 +344,23 @@ animals %>%
 animals %>% 
   tidyplot(x = diet, y = speed, color = family) %>% 
   add_barstack_relative()
+```
+
+# Statistical comparison
+
+```{r}
+
+```
 
+# Annotation
+
+```{r}
+
+```
+
+
+
+```{r}
 # curves
 
 energy_week %>%