Introduction to R

Introduction to R

  • In this course, we will review formulas, but we will use R for computational purposes.

    • Remember to refer to the lecture notes for specific code needed.
    • Code is also available on this course’s GitHub repository.

  • You can install R and RStudio if you wish; both are free.

  • We also have access to the Posit Workbench (“the server”) through HMCSE.

  • I know that this is probably the first time you are seeing R (or any sort of programming).

    • That is why we have “R lab” time built in to our course.
    • Remember that I am not looking for perfection, but for competency.

Introduction to R

  • Please download today’s activity from Canvas and log into the server.
    • Click on “New session”
    • Click on “Create session”
  • Upload today’s activity to the server:
    • In the bottom right pane, click on the white square with a golden up arrow
    • Click on “Choose file”
    • Navigate to the downloaded file
  • Open today’s activity on the server:
    • In the bottom right pane, scroll to the bottom
    • Click on the name of the .qmd file for today’s activity

Introduction to R: .R scripts

  • .R scripts:
    • Only allows code
      • Can comment out code using pound sign
    • Can run code line-by-line
    • Can run multiple lines of code at a time
    • Results output to Console window pane (bottom left)
  • I use .R scripts for my day-to-day analyses

Introduction to R: .qmd file

  • .qmd files:
    • Allow both text and code
      • Can comment out text using html code: <!-- comment here -->
      • Can comment out text in code chunk using pound sign: # comment here
    • Uses “code chunks” to evaluate code
      • Button: run all chunks before
      • Button: run this chunk
      • Ctrl+enter / cmd+return: line-by-line
    • Rendering results in .html file
  • I use .qmd files to create sharable documents
    • Reproducible research forever and always

Introduction to R: Disclaimer!

  • My major disclaimer as a biostatistician: I am a statistician first, programmer second.
    • My expertise is in statistics, not programming.
    • I do not know everything about R.
    • I do not claim to write the most efficient code.
    • Our goal is to correctly apply statistics to answer research questions using data.
      • R is a tool for us to apply statistics.
  • My major disclaimer as a professor: yes, I know this is likely the first time you are seeing R or programming in general.
    • We have “R lab” time built into the course.
    • Code you need to answer required questions will always be provided in lecture.
      • This means you must revisit lecture slides to find the code you need.

Introduction to R: Dr. Seals’s Expectations

  • I expect students to try their best. This includes:

    • referring back to lectures as needed.
    • asking when you have a question.
    • using the resources provided to learn.

  • You must know how to answer questions using R.

  • You will not be expected to write code beyond what is shown in class.

    • Note: Sometimes I include bonus questions…

  • When grading, I am looking for competency.

    • What is the appropriate analysis for the question at hand?
    • What are the assumptions of the analysis? Do we meet them?
    • Are the correct conclusions drawn given the information provided?

Functions in R: base R vs. packages

  • R functions are like baking recipes. They:
    • Take input (ingredients, or data),
    • Does something with it (follow recipe, or perform calculations),
    • Gives back a result (baked good, or statistics).
  • Some functions in R are available as soon as you open RStudio (this is “base R”).
    • e.g., mean(), sd()
  • Other functions are not available and must be called in after you start RStudio (these are “packages”).
    • e.g., after I call in library(tidyverse), I can use summarize(mean(), sd())
    • We will always need library(tidyverse) because of %>% (pipe operator).

Functions in R: tidyverse

  • library(tidyverse) is a collection of R packages designed for data science.
    • All packages share a common philosophy and are meant to work together.
    • This is ideal for using the “same syntax” - I promise it’s better than base R!
  • Core library(tidyverse) packages we will use:
    • library(readr): read in data files
    • library(dplyr): manipulate and summarize data
    • library(ggplot2): create data visualizations
  • If you are interested, there are resources:

Functions in R: ssstats

  • library(ssstats) is the package I have developed for this course.
    • I have made all of the R syntax consistent across functions.
    • Everything is also tidyverse friendly (ready for %>%).
    • Goal: focus on learning concepts.
  • This is our second semester piloting the package.
    • There are probably going to be bumps in the road.
    • Let me know when you encounter issues and I will help you determine if it’s user error or package error.

Functions in R: Summarizing Continuous Data

  • We will use mean_median() from library(ssstats) to summarize continuous variables.
    • It will return both the mean (standard deviation) and median (IQR).
dataset_name %>% 
  mean_median(var1, var2, ...)
  • We can add group_by() from library(tidyverse) to split the summaries by categories.
dataset_name %>% 
  group_by(grouping_var1, grouping_var2, ...) %>% 
  mean_median(var1, var2, ...)

Functions in R: Summarizing Continuous Data

  • Let’s use mean_median() to summarize the MLP dataset.
mlp_data %>% 
  mean_median(friendship, tail_shimmer, magical_energy)

Functions in R: Summarizing Continuous Data

  • Let’s use mean_median() to summarize the MLP dataset.
mlp_data %>% 
  mean_median(friendship, tail_shimmer, magical_energy)
# A tibble: 3 × 3
  variable       mean_sd      median_iqr   
  <chr>          <chr>        <chr>        
1 friendship     7.6 (1.6)    8.0 (2.0)    
2 magical_energy 9.9 (9.6)    7.0 (11.1)   
3 tail_shimmer   256.6 (65.7) 253.0 (103.0)

Functions in R: Summarizing Continuous Data

  • Let’s use mean_median() to summarize the MLP dataset by pony type.
mlp_data %>% 
  group_by(type) %>%
  mean_median(friendship, tail_shimmer, magical_energy)

Functions in R: Summarizing Continuous Data

  • Let’s use mean_median() to summarize the MLP dataset by pony type.
mlp_data %>% 
  group_by(type) %>%
  mean_median(friendship, tail_shimmer, magical_energy)
# A tibble: 12 × 4
   type    variable       mean_sd      median_iqr   
   <chr>   <chr>          <chr>        <chr>        
 1 Alicorn friendship     7.8 (1.3)    8.0 (2.0)    
 2 Earth   friendship     7.6 (1.6)    8.0 (2.0)    
 3 Pegasus friendship     7.6 (1.6)    8.0 (2.0)    
 4 Unicorn friendship     7.5 (1.6)    8.0 (2.0)    
 5 Alicorn magical_energy 9.0 (8.0)    6.2 (10.9)   
 6 Earth   magical_energy NaN (NA)     NA (NA)      
 7 Pegasus magical_energy NaN (NA)     NA (NA)      
 8 Unicorn magical_energy 9.9 (9.6)    7.0 (11.1)   
 9 Alicorn tail_shimmer   280.1 (64.5) 297.0 (104.0)
10 Earth   tail_shimmer   252.2 (65.2) 246.0 (100.0)
11 Pegasus tail_shimmer   263.5 (67.0) 265.0 (110.0)
12 Unicorn tail_shimmer   261.2 (65.0) 260.0 (100.0)

Functions in R: Summarizing Categorical Data

  • We will use n_pct() from library(ssstats) to summarize categorical variables.

  • For one variable – this returns n_i \ (\%_i):

dataset_name %>% 
  n_pct(var1)
  • For two variables – this returns n_{ij} \ (\%_{\text{col}}):
dataset_name %>% 
  n_pct(var1, var2)

Functions in R: Summarizing Categorical Data

  • Let’s use n_pct() to summarize the MLP dataset.
mlp_data %>% 
  n_pct(type, rows = 4)

Functions in R: Summarizing Categorical Data

  • Let’s use n_pct() to summarize the MLP dataset.
mlp_data %>% 
  n_pct(type, rows = 4)
    type      n (pct)
 Alicorn    41 (1.4%)
   Earth 1678 (58.4%)
 Pegasus  487 (17.0%)
 Unicorn  665 (23.2%)

Functions in R: Summarizing Categorical Data

  • Let’s use n_pct() to summarize the MLP dataset.
mlp_data %>% 
  n_pct(friendship, type, rows = 4)

Functions in R: Summarizing Categorical Data

  • Let’s use n_pct() to summarize the MLP dataset.
mlp_data %>% 
  n_pct(friendship, type, rows = 4)
# A tibble: 4 × 5
  friendship Alicorn  Earth     Pegasus   Unicorn  
       <dbl> <chr>    <chr>     <chr>     <chr>    
1          1 0 (0.0%) 0 (0.0%)  0 (0.0%)  1 (0.2%) 
2          2 0 (0.0%) 5 (0.3%)  1 (0.2%)  1 (0.2%) 
3          3 0 (0.0%) 14 (0.8%) 6 (1.2%)  13 (2.0%)
4          4 2 (4.9%) 54 (3.2%) 14 (2.9%) 16 (2.4%)

Graphs in R: Using ggplot()

  • We will construct data visualizations using library(ggplot2), which loads in when we load library(tidyverse).

  • This package allows us to create a layered visualization.

    • ggplot() creates the base layer.
    • geom_X() creates the individual pieces.
      • geom_point() creates a scatterplot.
      • geom_line() creates connected lines.
      • geom_bar() creates a bar chart.
      • geom_histogram() creates a histogram.

Graphs in R: Using ggplot()

  • We use ggplot() because it is very flexible - it allows us to customize every part of the graph.

    • Note that customization is less important in this course, but incredibly important in real life.

  • The R Graphics Cookbook is a great place to get basic code for graphs.

  • Remember that I do not expect you to memorize code. I do not have the code memorized.

    • Things I regularly ask Google for help with:
      • How to suppress the legend.
      • How to specify the tickmarks on the axis.
      • How to change the font size.

Graphs in R: The ggplot() Layer

  • Calling ggplot() creates the initial layer the graph lasagna.
mlp_data %>% ggplot()

Graphs in R: The ggplot() Layer

  • We specify the aesthetics through aes() in ggplot().
mlp_data %>% ggplot(aes(x = tail_shimmer, y = flying_speed))

Graphs in R: Overriding Defaults

  • We can override plot defaults using additional layers.
mlp_data %>% 
  ggplot(aes(x = tail_shimmer, y = flying_speed)) +
  labs(x = "Tail Shimmer",
       y = "Flying Speed") +
  theme_bw()

Graphs in R: Overriding Defaults

mlp_data %>% 
  ggplot(aes(x = tail_shimmer, y = flying_speed)) +
  labs(x = "Tail Shimmer",
       y = "Flying Speed") +
  theme_bw()

Graphs in R: Box Plots

  • Construct a box plot for the tail shimmer of the ponies (tail_shimmer).
mlp_data %>% ggplot(aes(x = tail_shimmer)) +
  geom_boxplot() +
  labs(x = "Tail Shimmer") +
  theme_bw() +
  theme(axis.ticks.y = element_blank(),
        axis.text.y = element_blank())

Graphs in R: Box Plots

  • Construct a box plot for the tail shimmer of the ponies (tail_shimmer).

Graphs in R: Box Plots

  • Construct a box plot for the tail shimmer of the ponies (tail_shimmer).
mlp_data %>% ggplot(aes(y = tail_shimmer)) +
  geom_boxplot() +
  labs(y = "Tail Shimmer") +
  theme_bw() +
  theme(axis.ticks.x = element_blank(),
        axis.text.x = element_blank())

Graphs in R: Box Plots

  • Construct a box plot for the tail shimmer of the ponies (tail_shimmer).

Graphs in R: Histograms

  • Construct a histogram for the flying speed of ponies (flying_speed).
mlp_data %>% ggplot(aes(x = flying_speed)) +
  geom_histogram(bins = 15, 
                 color = "#2E7D32", 
                 fill = "#4CAF50") +
  labs(x = "Flying Speed", 
       y = "Number of Ponies") +
  theme_bw()

Graphs in R: Histograms

  • Describe the histogram of the flying speed of ponies (flying_speed):

Graphs in R: Histograms

  • Construct a histogram for the magical energy of ponies (magical_energy).
mlp_data %>% ggplot(aes(x = magical_energy)) +
  geom_histogram(bins = 15, 
                 color = "#8B6C42", 
                 fill = "#F0E9DD") +
  labs(x = "Magical Energy",
       y = "Number of Ponies") +
  theme_bw()

Graphs in R: Histograms

  • Describe the histogram of the magical energy of ponies (magical_energy):

Graphs in R: Bar Graphs

  • Construct a bar graph for the combined age and sex of ponies.
mlp_data %>%
  count(sex) %>%
  ggplot(aes(x = sex, y = n)) +
  geom_col() +
  labs(x = "Age and Sex of Pony",
       y = "Number of Ponies")+
  theme_bw()

Graphs in R: Bar Graphs

  • Construct a bar graph for the combined age and sex of ponies.

Graphs in R: Bar Graphs

  • Construct a bar graph for the type of pony.
mlp_data %>%
  count(type) %>%
  ggplot(aes(x = type, y = n)) +
  geom_col() +
  labs(x = "Type of Pony",
       y = "Number of Ponies")+
  theme_bw()

Graphs in R: Bar Graphs

  • Construct a bar graph for the type of pony.

Graphs in R: Scatterplots

  • Construct a scatterplot with magical energy (magical_energy) on the x-axis and tail shimmer (tail_shimmer) on the y-axis.
mlp_data %>% ggplot(aes(y = tail_shimmer, x = magical_energy)) +
  geom_point(size = 2) +
  labs(x = "Magical Energy",
       y = "Tail Shimmer") +
  theme_bw()

Graphs in R: Scatterplots

  • Construct a scatterplot with magical energy (magical_energy) on the x-axis and tail shimmer (tail_shimmer) on the y-axis.

Graphs in R: Scatterplots

  • Construct a scatterplot with magical energy (magical_energy) on the x-axis and flying speed (flying_speed) on the y-axis.
mlp_data %>% ggplot(aes(x = magical_energy, y = flying_speed)) +
  geom_point(size = 2) +
  labs(y = "Tail Shimmer",
       x = "Flying Speed (km/h)") +
  theme_bw()

Graphs in R: Scatterplots

  • Construct a scatterplot with magical energy (magical_energy) on the x-axis and flying speed (flying_speed) on the y-axis.

Wrap Up

  • Always remember that I do not expect you to:
    • Memorize code.
    • Produce code in a timed environment.
    • Automatically know how to do these things.
  • I do expect you to:
    • Use your resources (lecture slides, GitHub website, Discord).
    • Try your best.