Learning Objectives

After this topic, students will be able to:

• Install packages for data visualization in R
• Use ggplot plotting functions
• Add plots to an rmarkdown document
• Explain what metadata is

---

Day 1

Class set up (5 min)

• Open your data-science RStudio project
• make sure that everything is up to date from the remote and that you have pushed all previous changes.
• make sure that your working directory is set up correctly with getwd and setwd
• Make sure that you have a “documents” folder. All your Rmd files will go in there.

Intro to the working dataset and class set up (15 min)

• For the data visualization class, we will be working with the “UHURU” dataset, that was published in 2014 in the journal of Ecology, in a paper entitled Plant and small-mammal responses to large-herbivore exclusion in an African savanna: five years of the UHURU experiment.
• All information about the UHURU data set is available on its metadata description.
• Go to the images available in the metadata description.
  • Instructor verbally describes the experiment and its goal using the images
• Authors made the raw data itself available on figshare.
• There are several data files associated to this publication, that can be downloaded from figshare.
• For this lecture, we will only work with the ACACIA_DREPANOLOBIUM_SURVEY.txt data file. Click on the name of the file to download it to your computer.
• Acacia drepanolobium is one of aprox 160 species in the genus Acacia. These are trees that live in dry forests, and they are very common in Mexico. They have yellow bright, pom-pom liked flowers, and have a symbiotic relationship with ants. Look at some images of their flowers on google.
• In your data-science RStudio project create a new Rmd file.
• Save the new Rmd file in the “documents” folder with the name “uhuru-dataset-visualization.Rmd”
• Use markdown syntax to start a tiny description of the dataset that we will be using during class
• Finding out the working directory of our R chunks

Reading TSV files (10 min)

• To start working with the dataset, we need to read it into R and create an object.
• Which function do we use for that?
  • read.csv()
• Use read,csv() to read the file “ACACIA_DREPANOLOBIUM_SURVEY.txt” that you saved in the “data-raw” folder
  • run read.csv(file = "../data-raw/ACACIA_DREPANOLOBIUM_SURVEY.txt")
  • if a “read error” comes up, try the following:
    1. make sure that your data set file is in the “data-raw” folder
    2. make sure there aren’t any typos in the file name
    3. make sure that your Rmd file is in the “documents” folder.
  • if the “error in connection” or “error reading file persists, try using the absolute path of the file for now. But keep in mind:
    • ⚠️ Avoid using absolute paths as much as possible!
    • Using relative paths is one of the best practices for good data science.
• When the code runs correctly, what does the output look like?
  • we do not get a table
  • values are separated by the characters “\t” and not commas “,”
• “\t” is how we indicate a Tab character in programming
• the data set is a “tab separated values” or TSV file.
• read.csv() read CSV files by default
• to read TSV files we need to use the argument sep = in the read.csv() function
• run read.csv(file = "../data-raw/ACACIA_DREPANOLOBIUM_SURVEY.txt", sep = "\t")
  • does the output look like a table now?
• assign the output to an object called acacia
  • run acacia <- read.csv(file = "../data-raw/ACACIA_DREPANOLOBIUM_SURVEY.txt", sep = "\t")

Data quality control check (10 min)

• what functions can we use to visually explore the structure and contents of our data set?
  • run head(acacia) to see the first 6 rows and all columns
  • run summary(acacia) to see a summary of data from each column
  • run View(acacia) to see the whole table
• How do we know the class of our variables?
  • run class(acacia) and typeof(acacia)
  • run class() for each column in acacia
  • run sapply(acacia, class)
• Identify missing/incorrect values
• Use the na.string = option from the read.csv() function to read the data set again and assign NA to missing/incorrect values
• run acacia <- read.csv(file = "/Users/lunasare/Desktop/data-science-course/fall-2022/data-raw/ACACIA_DREPANOLOBIUM_SURVEY.txt", sep = "\t", na.strings = "dead")
• Make sure that the “HEIGHT” column is numeric
  • run class(acacia$HEIGHT)
  • run class(acacia$HEIGHT) == "numeric"
  • run is.numeric(acacia$HEIGHT)

Intro to the package ggplot2 (10 min)

• “ggplot” stands for “grammar of graphics plot”
• uses a more “natural” way of creating plots by describing what you want, not how to build it
• we construct graphs by addibg layers of graphical elements, instead of modifying the whole code
• install the package
• load the package
• the function ggplot() creates a base “ggplot” object that we can then add things to
• it works like a blank canvas
• The two main arguments we will use for the function ggplot() are:
  • data = assigns the the name of the data frame we want to plot
  • mapping = assign columns from the data that we want to plot on each axis
• the aes() function is used within the option mapping =
  • For example, to plot the relationship between the circumference and the height of the trees, we assign the circumference column “CIRC” to the x axis and the height column “HEIGHT” to the y axis as follows:

    ggplot(data = acacia, mapping = aes(x = CIRC, y = HEIGHT))
    

A scatter plot (10 min)

• Once we haave the base layer, we can start adding points
  1. Then, use the ` + ` sign to add the next layer
  2. It is standard to hit “Enter” after the plus so that each layer shows up on its own line
  3. to add a scatter plot we use the function geom_point()

     ggplot(data = acacia, mapping = aes(x = CIRC, y = HEIGHT)) +
     geom_point()
     

• We can change the look of the scatter plot using arguments from the geom_point() function:
  • the shae of the points is determined by the shape = argument.
  • the size of the points is determined by the argument size = .
    • Small numbers like 1 or 2 will plot small points.
    • Large numbers like 10 or 11 will plot larger points.
    • You can give decimal numbers, like 0.1 to plot really small points, or give really large numbers, like 150, to plot super big points.
  • the color of the points is determined by the argument color = .
    • Colors should be written as a character vector. For example, "blue" will plot blue points.
    • For a list of in-built R colors, run the function colors()
  • the transparency of the points, which is called alpha, is determined by the argument alpha =
    • transparency goes from 0 to 1, 0 being the most transparent and 1 being the most solid.
    • a transparency of 0.5 will plot points that have a medium transparency.

      ggplot(data = acacia, mapping = aes(x = CIRC, y = HEIGHT)) +
      geom_point(size = 3, color = "blue", alpha = 0.5)
      

Adding labels (5 min)

• to add labels and a title to our plot, we use the labs() function:

  ggplot(data = acacia, mapping = aes(x = CIRC, y = HEIGHT)) +
  geom_point(size = 3, color = "blue", alpha = 0.5) +
  labs(x = "Circumference", y = "Height",
       title = "Acacia Survey at UHURU")
  

Rescaling axes (5 min)

• We can transform the scale of the axes in many ways
• This does not change the data itself, just the presentation of it
• to do a log 10 rescale of our axis, we use the functions scale_x_log10() and scale_y_log10()

  ggplot(data = acacia, mapping = aes(x = CIRC, y = HEIGHT)) +
  geom_point(size = 3, color = "blue", alpha = 0.5) +
  scale_y_log10() +
  scale_x_log10()
  

Grouping: Plotting data by another variable (10 min)

• For example, what if we want to see the effect of the experimental treatment (stored in the column “TREATMENT”)
• we have two main ways to group data by a third variable:
  1. plot data on the same graph, but using different colors for each value from the third variable
     • we will remove the argument color = from geom_point()
     • then we will add the argument color = to our first layer, inside ggplot()
     • there we will provide the name of the third variable as option:

       ggplot(acacia, aes(x = CIRC, y = HEIGHT, color = TREATMENT)) +
       geom_point(size = 3, alpha = 0.5)
       

  2. plot data on a subplot for each value from the third variable, with no color:

     ggplot(acacia, aes(x = CIRC, y = HEIGHT)) +
     geom_point(size = 3, alpha = 0.5) +
     facet_wrap(~TREATMENT)
     

     or with color:

     ggplot(acacia, aes(x = CIRC, y = HEIGHT, color = TREATMENT)) +
       geom_point(size = 3, alpha = 0.5) +
       facet_wrap(~TREATMENT)
     

In-class Activity (10 min)

• The UHURU data set

---

---

Day 2

Introduction (30 min)

• We are working on the Rmd file “uhuru-dataset-visualization.Rmd”
• all code and writing should be added there
• reviewing and explaining stuff I missed from day 1
  • working directory in an r chunk is not the same as a project working directory
  • change absolute paths to relative paths
  • shape argument in geom_point(),
  • the function color()
  • shape numbers with ?pch
  • subplots with facet_wrap()

In-class Activity (5 min)

• Add some interpretations to your plots
• render to PDF

Fitting linear models (10 min)

• the geom_smooth() function allows finding trends in our data

  ggplot(acacia, aes(x = CIRC, y = HEIGHT)) +
  geom_point(size = 3, alpha = 0.5) +
  geom_smooth()
  

• we can group by treatment and fit a linear model to each grouped
• what happens when we fit a linear model to data grouped by treatment?

  ggplot(acacia, aes(x = CIRC, y = HEIGHT, color = TREATMENT)) +
  geom_point(size = 3, alpha = 0.5) +
  geom_smooth()
  

• use the argument linetype = to assign different lines by treatment

  ggplot(acacia, aes(x = CIRC, y = HEIGHT, color = TREATMENT, linetype = TREATMENT)) +
  geom_point(size = 3, alpha = 0.5) +
  geom_smooth()
  

• use the argument method = from the function geom_smooth() to assign a specific method for model fitting

  ggplot(acacia, aes(x = CIRC, y = HEIGHT, color = TREATMENT)) +
  geom_point(size = 3, alpha = 0.5) +
  geom_smooth(method = "lm") # try with "glm"
  

Data distribution: Histograms and bar plots (10 min)

• Function geom_bar() creates bar plots and can be used for categoric or discrete variables
• Create a bar plot of number of plants for each treatment:

  ggplot(data = acacia, mapping = aes(x = TREATMENT)) +
  geom_bar()
  

• Function geom_histogram() creates histograms and can only be used for continuous variables
• you can define the number of bins with bins = and the bin width with binwdth =
• the fill = argument allows defining the color inside the bars
• Let’s show the distributon of circumference values with red colored bars:

  ggplot(acacia, aes(x = CIRC)) +
  geom_histogram(fill = "red")
  

• Now show the proportion of data colored by treatment

  ggplot(acacia, aes(x = CIRC, color = TREATMENT)) +
  geom_histogram(bins = 20)
  

• Or use the treatment data to assign a fill color with a gray outline:

  ggplot(acacia, aes(x = CIRC, fill = TREATMENT)) +
  geom_histogram(bins = 20, color = "gray")
  

Layering multiple data (10 min)

• To plot different data or variables on the same plot, we can change these values across layers

  ggplot() +
  geom_point(data = acacia,
             mapping = aes(x = CIRC, y = HEIGHT,
                           color = TREATMENT)) +
  geom_smooth(data = acacia,
              mapping = aes(x = CIRC, y = HEIGHT))
  

• we can even plot categorical and continuous data on the same plot:

  ggplot() +
  geom_point(data = acacia,
             mapping = aes(x = CIRC, y = HEIGHT,
                           color = TREATMENT)) +
  geom_smooth(data = acacia,
              mapping = aes(x = CIRC, y = HEIGHT)) +
  geom_histogram(data = acacia,
                 mapping = aes(x = CIRC, color = TREATMENT),
                 alpha = 0.1)
  

• We can plot data from two different columns as Y or X, by adding a layer and changing the corresponding value in the aesthetic
• We can use data from two different data sets by changing the value of data =
• as you can see we can combine all tools and use them in any “geom_” function in different ways!

Saving plots as image files (5 min)

• We will use the function ggsave()
• filename = is its main argument. It indicates:
  • the path where the image will be saved
  • the name that we want to give to the saved image
  • the type of file we will save given as an extension (.pdf, .jpg, .png)
• You run this function after you have run all the code to create your plot

  ggsave("acacia_by_treatment.jpg")
  

• other arguments that you may use control the size of the image in inches:

  ggsave("acacia_by_treatment.pdf", height = 5, width = 5)
  

In-class Activit (20 min) and home exercises (20 min)

• Acacia and ants