Data visualization - exercise sheet

Block 1: preparation and scatter plots

1.1 ggplot2 is part of the tidyverse package collection, so we’ll need to load it:

library(tidyverse)

1.2 Then import a version of the surveys data set that we worked with previously (surveys_complete.csv). Assign it to a tibble called surveys_complete, using read_csv(). We can use a version of the data located in the /shared folder.

surveys_complete <- read_csv("/home/rstudio/shared/surveys_complete.csv")

1.3 Using ggplot2 and the surveys_complete data set, create a scatter plot with weight on the y axis and species_id on the x axis, with the plot types shown in different colors. What do you think - is this a good way to show these type of data? If not, how would you go about it instead?

ggplot(data = surveys_complete, 
              mapping = aes(x = species_id, y = weight)) +
       geom_point(aes(color = plot_type))

1.4 Try setting up a plot template using the code you just wrote for creating the scatter plot above.

Just as a memory aid, here is a (non-exhaustive) list of different geoms we can use.

# geom_point() for scatter plots, dot plots, etc.
# geom_bar() for bar plots
# geom_boxplot() for box plots
# geom_line() for trend lines, time series, etc.  

# Assign plot to a variable
template <- ggplot(data = surveys_complete,
                   mapping = aes(x = species_id, y = weight))

# Draw the plot
template + 
    geom_point(aes(color = plot_type))

Block 2: bar plots and custom colour palette

2.1 Recreate the the bar plot we were just examining, showing mean weights of each species. This time, also try adding the following custom palette to the plot (with the data colour by species_id). The palette contains a list of different RGB values that can be used for creating color-blind plots.

cPalette <- c("#000000","#004949","#009292","#ff6db6","#ffb6db",
              "#490092","#006ddb","#b66dff","#6db6ff","#b6dbff",
              "#920000","#924900","#db6d00","#24ff24","#ffff6d")

To employ this custom palette, one would use one of the following:

scale_fill_manual(values=cPalette) # for fills
scale_colour_manual(values=cPalette)  # for lines and points

meanwt <- surveys_complete |>
           summarize(mean_weight = mean(weight),
           .by = species_id)
           
ggplot(data = meanwt, mapping = aes(x = species_id,
                                    y = mean_weight)) +
geom_col(aes(fill = species_id)) +
scale_fill_manual(values = cPalette)

Block 3: histograms

3.1 Create a histogram that compares the distribution of weight values between males and females in the surveys_complete data set. You should be able to tell apart the two sexes based on how the bars are colored. For this exercise, use a white fill, a binwidth of 4 and an alpha value of 0.5.

Tip: When plotting two data sets together in the same histogram, adding position = "dodge" inside the geom can make the results easier to interpret (it asks R to produce an interleaved plot, i.e. values are shown side-by-side rather than stacked on top of one another).

ggplot(surveys_complete, aes(x = weight, color = sex)) +
  geom_histogram(fill = "white", alpha = 0.5, binwidth = 4, position = "dodge") 

Block 4: box and violin plots

4.1 Box plots are useful summaries, but hide the shape of the distribution. For example, if the distribution is bimodal, we would not see it in a box plot. In addition to histograms, an alternative to the box plot is the violin plot, where the shape (of the density of points) is drawn.

Try modifying the following code to create a violin plot instead (you can use geom_violin()). Can you think of anything else that could be done to improve the presentation of the data?

ggplot(data = surveys_complete, mapping = aes(x = species_id, y = weight)) +
    geom_boxplot(alpha = 0)

ggplot(data = surveys_complete, mapping = aes(x = species_id, y = weight)) +
  geom_violin(alpha = 0)

One could change the y axis scale (a linear scale doesn’t seem to work that well).

4.2 In many types of data, it is important to consider the scale of the observations. For example, it may be worth changing the scale of the axis to better distribute the observations in the space of the plot. Changing the scale of the axes is done similarly to adding / modifying other components (i.e., by incrementally adding commands).

Using the code you just created for the violin plot above, represent weight on the log10 scale instead; see scale_y_log10().

ggplot(data = surveys_complete, mapping = aes(x = species_id, y = weight)) +
  geom_violin(alpha = 0) +
  scale_y_log10()

4.3 Create a boxplot for hindfoot_length. Overlay the boxplot layer on a jitter layer to show actual measurements.

ggplot(data = surveys_complete, mapping = aes(x = species_id, y = hindfoot_length)) +
geom_jitter(alpha = 0.3, color = "tomato") +
geom_boxplot(alpha = 0) 

Block 5: line plots and some preparation

Take some time to construct the yearly_counts data set for yourself, and trying out a line plot.

5.1 The yearly_counts object should contain numbers of counts per year for each genus. We need to group the data and count records within each group:

yearly_counts <- surveys_complete |>
  count(year, genus)

5.2 Create a line plot of the yearly_counts data as follows. Remember that we can use the color argument does a nice job at grouping the data for us.

ggplot(data = yearly_counts, mapping = aes(x = year, y = n, color = genus)) +
    geom_line()

5.3 Let’s create another object, called yearly_sex_counts. This should be similar to yearly_counts, but with the data additionally grouped by sex. We will need that in later exercises!

yearly_sex_counts <- surveys_complete |>
  count(year, genus, sex)

Block 6: faceting

Here’s a reminder of the two options we have for faceting using ggplot2:

• facet_wrap() arranges a one-dimensional sequence of panels.
• facet_grid() allows you to form a matrix of rows and columns of panels.

We’ve tried facet_wrap(). Next, let’s check out facet_grid().

6.1 You might remember this facet_wrap() example that we covered earlier:

ggplot(data = yearly_sex_counts, mapping = aes(x = year, y = n, color = sex)) +
  geom_line() +
  facet_wrap(facets =  vars(genus))

Try using facet_grid() to control how panels are organised by both rows and columns, with rows corresponding to sex and columns to genus. Instead of facets = vars(genus), we would now have to use rows = vars(...) and cols = vars(...). We haven’t yet gone through the code formatting for this, so don’t worry if you can’t get it right!

ggplot(data = yearly_sex_counts, 
       mapping = aes(x = year, y = n, color = sex)) +
  geom_line() +
  facet_grid(rows = vars(sex), cols = vars(genus))

6.2 We can also use facet_grid() to organise the panels only by rows (or only by columns). See if you can figure out both options.

# One column, facet by rows
ggplot(data = yearly_sex_counts, 
       mapping = aes(x = year, y = n, color = sex)) +
  geom_line() +
  facet_grid(rows = vars(genus))
  
# One row, facet by column
ggplot(data = yearly_sex_counts, 
       mapping = aes(x = year, y = n, color = sex)) +
  geom_line() +
  facet_grid(cols = vars(genus))

Block 7: further customization

7.1 Try creating your own custom theme (you could call it mytheme) and using it to modify the appearance of the following plot. There’s no right or wrong answer here, feel free to experiment!

ggplot(surveys_complete, aes(x = species_id, 
y = hindfoot_length)) +
    geom_boxplot() +
    mytheme

One example of a theme can be found here:

grey_theme <- theme(axis.text.x = element_text(colour = "grey20",  
                    size = 12, angle = 90, hjust = 0.5, 
                    vjust = 0.5),
                    axis.text.y = element_text(colour = "grey20",  
                    size = 12),
                    text = element_text(size = 16))

# Own example: a simple black and white theme

mytheme <- theme(
  panel.border = element_blank(),             # remove border
  panel.grid.major = element_blank(),         # remove grid lines
  panel.grid.minor = element_blank(),
  panel.background = element_blank(),         # remove background
  axis.line = element_line(colour = "black")  # add axis line
  )

ggplot(surveys_complete, aes(x = species_id, 
y = hindfoot_length)) +
    geom_boxplot() +
    mytheme

Block 8: arranging, saving and exporting

8.1 If you don’t have it already, create a folder called fig_output in your working directory.

8.2 Run through this rather big block of code to create a couple of plots that we were looking at earlier.

# plot 1
spp_weight_boxplot <- ggplot(data = surveys_complete, 
                             mapping = aes(x = genus, 
                             y = weight)) +
  geom_boxplot() +
  xlab("Genus") + ylab("Weight (g)") +
  scale_y_log10() +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))

# plot 2
spp_count_plot <- ggplot(data = yearly_counts, 
                         mapping = aes(x = year, y = n, 
                         color = genus)) +
  geom_line() + 
  xlab("Year") + ylab("Abundance")

8.3 We can use ggsave() to save the plots. Try saving a single plot first:

ggsave("fig_output/plot1.png", my_plot, 
width = 15, height = 10)

8.4 Next we could save a grid.arrange() plot. First use grid.arrange() to combine plots 1 and 2, then use ggsave() to save the result (use a DPI value of 100).

# if you don't have gridExtra loaded, then run: library(gridExtra)

plots_arranged <- grid.arrange(spp_weight_boxplot, spp_count_plot, 
ncol = 2, widths = c(4, 6))

ggsave("fig_output/comboplot.png", plots_arranged, 
width = 10, dpi = 100)

8.5 Now it’s time to save and export everything we’ve created so far. Can you remember how to do this? If not, you can revisit the instructions in the Starting with Data exercise sheet (steps 7.3 and 7.4).