Introduction to Statistical Computing

👨‍💻 Eugene Hickey @ Atlantic Technological University 👨‍💻

  • eugene.hickey@tudublin.ie
  • @eugene100hickey
  • github.com/eugene100hickey
  • www.fizzics.ie

Week One

  • Welcome to the course on statistical computing.

  • We will discuss how to make the most of your data.

  • Complimentary to the qualitative analysis you do with Emma.

  • Goal here is to provides skills in data analysis.

  • We’ll learn to use a software system called R.

Acknowledgments

  • Ruth Moran, co-pilot for this workshop Graduate Research & Education Training Officer.

  • slides produced using quarto

  • assessments produced using Rexams

Target Audience

  • Graduate students looking for better ways to present their data.

  • People currently using tools like MS Excel for analysis.

Why R?

  • Working with a mouse isn’t reproducible

    • difficult to log exactly what you’ve done
    • hard to repeat for a series of analyses
    • difficult to be inspired by other people’s work

  • Good to separate sources of data from the analysis

  • R uses series of commands that input, manipulate, and display data

  • Lots of contributors around the globe, diverse fields

    • this is the world of open source software

Course Outline

  1. Introduction to Statistical Computing & R - Saturday April 5th

  2. Statistical Inference - Saturday April 12th

  3. Manipulating Data - Saturday April 26th

  4. Getting and Cleaning Data - Saturday May ??

  5. Data Visualisation - Saturday June ??

Stuff we won’t be doing

  1. creating functions, iterating

  2. R is cracker at machine learning

  3. also great at making documents and presentations using quarto

  4. and websites and blogs

Assessment

  • Weekly online quizzes
    • get three attempts
    • best mark from the three
    • some multiple choice, some open answer
    • based on this weeks lecture
  • Exercises to be done within R
    • these are called swirl()
    • email results to me (screenshot of final page is fine)
  • Project at end of course
    • somewhat optional, can either do one on Emma’s section or on this part, or a combination
    • best to do something associated with your research that can be useful beyond this module

Resources

Books

Websites

Blogs and Podcasts

Online Courses

Miscellaneous

Installing R and RStudio

  • first R from CRAN

    • R is the engine

  • then RStudio from Posit

    • RStudio is the cockpit

  • alternative is to make an account at Posit Cloud

    • give 25 hours per month

  • R is case sensitive

  • This is a nice tutorial suite to explain installing R and RStudio:

RStudio Screen

Using RStudio

  • toolbar across the top
    • I don’t use this very much
  • set of quick links below that
    • top left (green plus sign) is about the only one I use
  • 4 Panes
    • top left for files or looking at data
    • bottom left for the console
    • top right for Environment - tells what variables are stored
    • bottom right for plots and help

Using RStudio (continued)

  • usual work flow is:
    • try commands out at the console (bottom left)
    • when that works, store them in a file (top left)
    • when sequence of commands works, put them into a document (also top left)

Extending R

  • installing R just gives you base R
  • beauty of this tool lies with packages 📦
  • we’ll look at installing these from three sources:
    • CRAN
    • Bioconductor
    • github

CRAN

  • CRAN
    • example, on console type install.packages("palmerpenguins")
    • this installs the palmerpenguins package (we’ll need this for this week’s quiz)
    • over 20k packages on CRAN (see list here)
    • sometimes esoteric (engsoccerdata)
    • sometimes cutting edge (deep learning)
    • each package heavily curated and maintained
    • Task Views good place to start looking

Bioconductor

  • Bioconductor
    • set of bioinformatics packages (lots of genomics)
    • start with install.packages("BiocManager")
    • then BiocManager::install("some_genomics_package") to use
    • list of packages here
    • about 3,000 packages, including genome builds

Github

  • github
    • packages in development
    • start with install.packages("devtools")
    • then devtools::install_github("developer_name/package_name")
    • almost 80k packages here
    • the package githubinstall is useful to search these

Installing Packages

  • try install.packages("scales")
    • will add functions from someone else’s work so you can use it
    • need to do this just once
    • to actually make this available, type library(scales)
    • do this every time you start R and want to use scales package
  • one package we’ll need is tidyverse, install.packages("tidyverse")
    • this might take a few minutes because such a large collection of packages
  1. Using the Console
  2. Storing Values
  3. Fundamental Data Types
  4. The Dataframe - Rows \(\times\) Columns
  5. Dataframe Columns
  6. Subsets of Dataframes
  7. Small Useful Functions

Using the Console

  • R is a calculator
  • can type, say, 45 + 17 and get the answer back
  • * is multiply, / is divide
  • get constants like pi for \(\pi\)
  • can run functions from the console like, say, sqrt(x=49) (or just sqrt(49))
  • can get help on these functions by typing, say, ?sqrt
  • to get list of help in the base package try base:: and then press tab

Storing Values

  • results of calculations can be stored
  • do this with my_square_root_result <- sqrt(49)
  • the <- reads as gets (can also us equals sign, =, but that’s sloppy)
  • keyboard short cut for <- is alt and - simultaneously
  • full list of RStudio shortcuts in Help on the toolbar (ironically, of course, there is a keyboard shortcut for keyboard shortcuts help)
  • list of stored values given in the Environment tab of RStudio

Fundamental Data Types

  • numeric (or double) is for numbers with decimals. Default for numbers.
  • integer for counting numbers. Type in x <- 72L to get integer 72
  • logical gives TRUE and FALSE
  • character gives text. Try typing x <- "I am Groot".
    • Equivalent is x <- 'I am Groot'
  • complex is for stuff like 14 + 3i. I’ve never used these. Ironically sqrt(-1) gives an error rather than i
  • factor. These deserve fuller explanation and get a slide of their own.

Factors

  • super useful when only limited number of possible values for a variable
  • examples like female / male or Alabama / Alaska / …. Wyoming
  • possible values are called levels
  • levels have an order, default is alphabetical but can adjust this
  • forcats package in the tidyverse deals with factors
  • have big impact on figures
    • figures will look different depending on whether a variable is a factor or a character
    • legend will be in the order of levels of a factor
  • can change between data types using the as. functions
    • e.g. as.character(), as.numeric()
    • check out all the function in forcats for dealing with factors
  • can figure out the data type using the class() function

Data Frames

  • these are the workhorse of R data types
  • look like spreadsheets
  • we’ll work with them a lot
  • organised by rows (across) and columns (down)
  • each column must have the same type
  • can examine data frames using the View() function (note, capital “V”)
  • can find column names using the names() function
  • related concepts are tibbles and data.tables

Subsets of Data Frames

  • can access individual column elements by specifying the row and column position
    • e.g. mtcars[2, 5] is the 2nd row of the fifth column
  • can access an entire row by leaving the column part blank
    • e.g. mtcars[2, ] is the whole 2nd row
  • likewise get the entire column by leaving the row slot empty
    • e.g. mtcars[,5] is the entire 5th column

Subsets of Data Frames

  • can also access columns using the dollar sign, $
    • e.g. mtcars$cyl gives the 2nd column, called cyl
    • note mtcars$cyl[2] gives the second element of this
  • can use a number range using : to get a bunch of values
    • e.g. mtcars[1:4, 2:4] takes a chunk of the mtcars data frame

Lists

  • dataframes have constraints

    • all columns must be the same type

    • rows and columns must be the same length

  • lists are more general

  • example, list_data <- list(c("Jan","Feb","Mar"), matrix(c(3,9,5,1,-2,8), nrow = 2), list("green",12.3))

  • list_data[[1]] gives the three months, etc

    • note the double square bracket,[[ ]]

Some Useful Functions

  • c() let’s you create a vector of quantities. Coerced to same type
  • is.na() to check if a missing value, sum(is.na()) gives total of these
  • dim() gives number of rows / columns in data frame, also length()
  • class() for nature of variable
  • summary(), str(), glimpse() show data frame parameters. Also skim() from the skimr package
  • I use %in% quite a bit, checks if value is among a bunch of values
    • "August" %in% month.name
    • "ÂŁ" %in% letters

Some Useful Functions

  • Sys.date(), Sys.time()
  • ls() gives list of variables
  • file.list() gives list of files
  • sessionInfo() tells you what packages are loaded
  • citation() tells you about a package
  • mean() calculates mean of a vector, sd() the standard deviation
  • complete.cases() returns TRUE if there are no missing values in a row

And Some Weird Characters

  • get to use a lot of the keyboard in R
    • we’ve seen $ to extract a column of a data frame
    • () are arguments to functions
    • [] are gateways to data frame elements
    • %>% is called the pipe and is really cool. Ctrl + Shift + M shortcut
    • some logical operators: ! is NOT, | (or ||) is OR, & (or &&) is AND

Workshop - Week One

Perform the Following Tasks:

  1. multiply the numbers \(25\times \pi\) and save the result to my_first_result

  2. make a vector of five numbers e.g. 34.1, 54.4, 71.5, 93.8, 22.6 and save them to my_second_result

  3. multiply my_second_result by 7 and print the results

  4. print the airquality dataset

  5. examine the airquality dataset using summary(), str(), and glimpse(). For the latter you’ll need the tidyverse library.

  6. store the values of Temp from the airquality dataset in a new variable you can call temperature

  7. calculate the mean() of temperature

  8. calculate the standard deviation, sd(), of temperature

  9. find the names of the columns in the Puromycin dataset

  10. look up the help for the USJudgeRatings dataset and find out what is meant by the “DECI” column name

  11. install and load up the package dslabs. Use is.na() and sum() to find the number of missing values in the us_contagious_diseases dataset

  12. what kind of variable is stored in olive$area?

  13. what are the levels of the factor in olive$region?

Assignments - Week One

  1. Complete week one moodle quiz

  2. Complete swirl() exercises

  • install.packages("swirl")

  • library(swirl)

  • install_course("R Programming E")

  • swirl()

  • choose course R Programming: The basics of programming in R

  • do the four exercises 1 (basic Building Blocks) to 4 (Vectors)

  • email the results to eugene.hickey@associate.atu.ie