Introduction to R and Positron

Carsten Lange

clange@cpp.edu

Cal Poly, Pomona

Learning Outcomes

Part 1: Setup

How to install R and Positron
What is the windows layout of Positron
How to use a (project) folder in Positron
How to extend R’s functionality with R-packages and which packages to install

Learning Outcomes

Part 2: How R Stores the Data:

Data types
Data objects in R

Learning Outcomes

Part 3: The tidyverse Package:

The Structure of R commands
About the tidyverse package for data frames
- select() and rename columns (variables)
- filter() rows (observations)
- mutate() (define columns (variables); overwrite old or create new)
- arrange() sort observations in a data frame.
- piping (connecting commands) with |>.

Part 1: Install and Setup R and Positron

A typical setup to work with R consists of two components:

the R Console which executes R code and
an integrated development environment (IDE) such as RStudio or Positron.

You can download R here: Download R

You can download Positron here: Download Positron

Note
- Install R before Positron
- If an older R version exists uninstall it before installing the newer version

RStudio — Integrated Development Environment (IDE)

(does not have newest features of a modern IDE)

RStudio Window

Video for First Steps to setup R and RStudio: Click here
(However, it is recommended to work wit Positron rather than RStudio )

Positron — Integrated Development Environment (IDE)

Positron Window

First steps to setup R and Posit can be found in this video: coming soon

Set Window Layout and Choose R Interpreter

RStudio Keybindings (needs to done only once)

Click Gear Icon -> Choose: Settings -> Search for: Keybindings ->
Toggle-On RStudio Keybindings

Always Work with Folders

Always open or create a folder first

The folder is the one where all your R ().r), Quarto (.qmd), and data (e.g., .csv) files are stored

Use Command Palette to Create a New `R` or `Quarto` file

CTRL SHIFT P (Windows) or ⌘ SHIFT P (Mac) to open Command Pallette:

Type either “New Quarto Document” or “New R Document” into search bar
Click and Create file
Save file right away

Open a File from an Existing Positron Folder

Positron: First Steps with an R File `(.r)`

Remember, always open a folder first!

An R file such as `MyFile.r` contains only R-code

Print “Hello world” using the print() command.
Assign values $3$ and $4$ to the legs a and b of an right-angled triangle.

calculate the hypotenuse $c$: \[c^2=a^2+b^2 \Longleftrightarrow c=\sqrt{a^2+b^2}\]
print the result using the cat(command)
Assign the number $2$ to the variable (“R” objects) a and run the cat() command again.

Try Positron with a Quarto File `(.qmd)`

Remember, always open a folder first!

A Quarto file such as `MyFile.qmd` contains text as well as R-code

Text is written in MarkDown
Code is surrounded by:

```{r}

MyCode goes here

```

Try Positron with a Quarto File `(.qmd)`

Instructions:

Print “Hello world” using the print() command.
Assign values $3$ and $4$ to the legs a and b of an right-angled triangle.

calculate the hypotenuse $c$: \[c^2=a^2+b^2 \Longleftrightarrow c=\sqrt{a^2+b^2}\]
print the result using the cat(command)

Note, now we want everything nicely commented!

WASM: R Runs in a Browser Including Libraries and Data

Print “Hello world” using the print() command.
Assign values $3$ and $4$ to the legs a and b of an right-angled triangle.
Calculate the hypotenuse $c$ \[c^2=a^2+b^2 \Longleftrightarrow c=\sqrt{a^2+b^2}\]
Print the result using the cat(command)
Assign the number $2$ to the variable (“R” object) a and run the cat() command again.

R Packages

R Packages extend R’s functionality. They have to be installed only once:

For example, to install the tidyverse package, type in the consol window:
install.packages("tidyverse)
Needs to be only done once!

After installation packages they need to be loaded in every new R script or Quarto file with: library().

Packages frequently used in this course (please install soon):

tidyverse: supports easy data processing .
rio: allows loading various data resources with one import() command from the user’s hard drive or the Internet.
janitor: provides functionality to clean data and rename variable names to avoid spaces and special characters.

Videos for the `rio` and the `tidyverse` Package

Example: How to install the tidyverse package: Click here

Video about the rio package: Click here

Using the `rio` and the `tidyverse` Package

Example: rio and tidyverse package (assuming they are installed already)

import() would not work if the rio package were not loaded.
select() would not work if the tidyverse package were not loaded.

Part 2: Data Types & Data Objects

Data Types: Which type of values can R store?
- numerical num (such as: 0.1, 2.3, 3.14157)
- numerical int (such as: 1, 2, 7)
- character chr (such as: “Hello”, “Hi”, “World”)
- categorical factor (such as: “Female”, “Male” Or: “small”, “medium”, “large”)
- boolean logic (True, False)

Data Objects: What are the containers R uses to store data?
- single value as: single entry
- list of entries as: vector
- table as: dataframe or tibble
- advanced objects can hold: plots, models, prediction results

Analogy: Data Types & Data Objects Example for Three Alcoholic Beverages

Data Types: Which type of fluids can we store?
- beer
- wine
- whiskey
Data Objects: What are the containers to store our liquids?
- bottles
- cartons (incl. six packs)
- cargo containers

Numerical Data Type (num and int): Numerical values (e.g., 1, 523, 7 or 3.45, 0.1, 8.0) are used for calculations. In contrast, ZIP-Codes are not numerical data type.

Character Data Type (chr): Storing sequence of characters, numbers, and/or symbols to form a word or even a sentence is called a character data type (e.g. first or last names, street addresses, or Zip-codes)

Categorical Data Type (factor): A factor is an R data type that stores categorical data in an effective way. factor data types are also required by many classification models in R.

Logic Data Type(logic): A data type that stores the logic states TRUE and FALSE is called a logic object (sometimes called Boolean)

Character Data Type (chr) a.k.a Labels:

Note that what is called a character in R is often called a string in other programming languages.

character data types must be surrounded by quotes:

Character variables can be concatenated with the cat() command:

Numerical Data Type (num): Numerical values are used for calculations (therefore ZIP-Codes are not numerical). The numerical data type is num and in some cases int (full number). In most of the cases you do not have to differentiate between int and num.

Categorical values are stored in R as a factor data type. A factor object signals certain R commands that the variable is categorical rather than character.
Note, some R commands recognize categorical variables automatically and a conversion from character to factor is not needed.

Sex is a character variable in the dataset People

Transforming the variable $Sex$ to a factor and looking at its structure (str()) again:

Logic variables: Store TRUE and FALSE. They can be combined with and/or. Internally True is stored as $1$ and False is stored as $0$

Print

Print Hello world! by using variable A:

Calculate with Variables and Output with `cat()`

A rectangular lot has a width of 200 feet (Width) and a length (Length) of 300 feet. Calculate the area (Area) and create a full sentence output.

Exercise: `cat()` command and Single Value Objects with Different Data Types

Assign your own first and last name, your ZIP code, and your your age, to three character variables (first name, last name, Zip code) and one numerical variable (age). Use var1, var2, var3, var4. Afterward, use Cat() to output a sentence like Carsten Lange is 55 years old and lives in ZIP code 92656 using the variables you had created.

Again: Data Types & Data Objects

Data Types: Which type of values can R store? ✓
- Numerical
- Character
- Categorical / Factor

Data Objects: What are the containers R uses to store data? ?

Data Objects

Single Value Object
Vector Object
Data Frame (Tibble) Object
List Object (not covered in this course)
Advanced Object such as plots, models, recipes

Single Value Object

Objects just store a single value:

Vector-Objects

A vector object stores a list of values (numerical, character, factor, or logic; mixing of data types is not allowed)

Example: Weather during the last three days in Stattown:

Vector objects can be used as arguments for an R command to calculate statistics such as the mean() or the number of entries in the vector (length()):

Data Frames (tibbles)

A data frame is similar to an Excel table. A data frame stores the values of R Vectors as variables entries in its columns .

Note, that the c() command combines values to a vector.

Below we show how the values from the four vectors VecDay, VecTemp, VecWindSpeed, and VevIsSunny are stored in the data frame DataWeather.

The columns hold the values from the four vectors and the rows (with the exception of the first row), hold the observations for the various days. The first row contains the variable names:

Data Frame from Titanic Data

Most of the times, we do not build a data frame from its vectors (columns). Instead we load the data frame from a file (for example, a csv file).

Below we load the Titanic dataset. Note, only the first six observations are shown.

We can see the structure of the data frame by using the str() command. This includes the type of all variables/vectors:

Extracting the Vectors and Performing Calculations (numerical Vectors)

Since the columns of a data frame are made up of vectors, we can extract these vectors, and use the values for data analysis (remember: observations are in the rows, variables are in the columns).

We can use the notation DataFrameName$VectorName to extract the vectors:

Extracting the Vectors and Performing Calculations (logical Vectors)

If we like, we can change a vector inside a data frame:

We can use the logical vector Survived (remember, TRUE=1, FALSE=0) to calculate the survival rate:

Summary Data Types

graph TD
VarGeneral["Data in General"]
VarChar["Labels/Char:<br/>- Last names<br/>- ZIP-codes<br/>- ID-numbers"]
VarCategory["Categorical/Factors<br/>limited number of variations:<br/>- Sex (M/F)<br/>- Grades (A, B, C, D, F)<br/>- Gender (M, F, NonBin)"]
VarNumeric["Numerical Data<br/>- Counts (# of accident)<br/>- Measures<br/>  (e.g., Weight=160.8)"]
VarNumericCon["Continuous/Num<br/>- Measures<br>  (e.g., Weight=160.8)"]
VarInt["Discrete/Int<br/>- Counts<br>  (e.g., # accidents)"]


VarGeneral --> VarChar
VarGeneral --> VarCategory
VarGeneral --> VarNumeric
VarNumeric --> VarInt
VarNumeric --> VarNumericCon


%% Define one reusable style
classDef ClassTopNodeCL text-align:center,color:#000000,fill:#ffffff,stroke:#000000,stroke-width:3.8px;

classDef ClassLowNodesCL text-align:left,color:#0000dd,fill:#dedea9,stroke:#ffffff,stroke-width:3.8px;

classDef ClassLowestNodesCL text-align:left,color:#0000dd,fill:#b3672b,stroke:#ffffff,stroke-width:3.8px;

%% Apply it to multiple nodes
class VarChar,VarNumeric,VarNumericCon,VarInt,VarCategory ClassLowNodesCL;
class VarGeneral ClassTopNodeCL;
class VarNumericCon,VarInt ClassLowestNodesCL

Summary Data Types and Objects

Data Type and Object Structure

Part 3: The `tidyverse` and Piping

Basics of R Commands

R commands consists of the command’s name followed by a pair of parentheses: command()

Inside the () we can define one or more arguments for the command.

Arguments in a command usually have names such as x= or data=
R does not require to use the argument’s name, but order matters
R commands have many arguments. Most have default values
We can nest commands. However, nesting too deeply makes code difficult to read.»

Structure of R Commands

Most R commands have the following structure: \[\begin{equation} \underbrace{DataNew}_{\text{R object storing the result}}= \underbrace{Command}_{\text{Name of the command}} \underbrace{(\overbrace{Data}^{\text{1. Argument: Data to process}}, \overbrace{Arg2, Arg3, \dots, ArgN}^{\text{More Arguments}})}_{\text{Arguments inside () and separated by comma}} \end{equation}\]

Often the data argument is the first argument in a command. Usually named data= or x=.»

Use a Command with and without Argument Names

All three examples are equivalent

Getting Help about a Command (e.g., `mean`)

Use ?mean or help(mean) in the RStudio console to see the default values.

You can also mark/highlight and then press F1

Try it for the mean() command.

Important Commands from `tidyverse`/`dplyr` Package

dplyr package is part of the tidyverse (meta) package

library(tidyverse) (loads the tidyverse and its packages)
select() selects columns (variables) from a data frame
filter() filters rows (observations) for specific criteria
mutate() calculates new or overwrites existing columns (variables) based on other columns (just like Excel)
arrange() sorts a data frame according to one or more columns in ascending order (use argument desc() for descending order)

Titanic Dataset

Example: Using the `tidyverse` for Data Analysis

Goal: Create a data frame with a few selected variables, that contains only female observations, and the fare in current U.S.-$.

The `select()` Command

select(DataMine, Var1, Var2) selects columns (variables) Var1 and Var2 from a data frame DataMine. The first argument is the data= argument followed by the names of the selected variables.
select(DataMine, -Var1, -Var2) selects all columns (variables) except Var1 and Var2 from a data frame DataMine.

Here is an example using the DataTitanic data frame with a few selected variables:

The `filter()` Command

The filter() command filters rows (observations) of a data frame for specific criteria. The first argument is the data= argument followed by the filter criteria.

E.g., filter for female passengers:
We use DataTitanicSelVar that we created in the previous slide at as a starting dataframe and save the result in DataTitanicSelVarFem.
Note, we have to use == instead of = for the criteria):

The `mutate()` Command

mutate() creates or overwrites columns (variables) based on other columns (just like Excel). The first argument is the data= argument followed by the instructions on how to create the new variable.

E.g., mutate calculates the FareIn2023Dollars by multiplying FareInPounds by $108.5$. The command uses DataTitanicSelVarFem from the previous slide:*

Summary

We now have a data frame with only women and columns $Survived$, $PasClass$, $Sex$, $Age$, and $FareIn2023Dollars$.

How did we get there:

We selected variables $Survived$, $PasClass$, $Sex$, $Age$, $FareInPounds$ and saved in DataTitanicSelVar
We filtered for females and saved in DataTitanicSelVarFem
We mutated to calculate a new variable $FareIn2023Dollars$ and saved finally in DataTitanicSelVarFemDolFare

Could this be done easier?

Note, overwriting data frames such as DataTitanic is usually a bad idea! Nesting the command is possible but very difficult to read.

Piping Schema

Alternative: Piping

(will be used throughout the course/book)

Shortcut for |>: CTRL SHIFT M (Windows) or ⌘ SHIFT M (Mac).

The pipe operator |> is for most practical purposes equivalent to %>%.

Why R?

Excel analytics is not reproducible
SPSS focuses on surveys
STATA and SAS are commercial products
- not free
- progress has to go through the corporate hierarchy and therefore is slower
- limited support community

R and/or Python

Analysis is always reproducible with little effort
free
extensive support
R or Python
- R is easier to understand for users with limited coding experience
- Python is faster in incorporating cutting-edge algorithms
- transfer from R to Python or vice versa is easy
- Quarto supports both R and Python even simultaneously in the same project

Python vs. R — The task

Let us compare code for the same tasks between R and Python:

Download the Titanic dataset
select the variables Sex, FareInPounds, Survived (renamed to: Surv)
Calculate a new column FareInDollars by multiplying FareInPounds by $108.5$
Filter for Sex being female
Calculate the mean of FareInDollars

Python vs. R — The Results (using pandas)

library(tidyverse)
library(rio)
DataTitanicR = import("Data/Titanic.csv") |>
  select(Sex, FareInPounds, Surv = Survived) |>
  mutate(FareInDollars = FareInPounds * 108.5) |>
  filter(Sex == "female")
MeanFareWomen = mean(DataTitanicR$FareInDollars)
print(MeanFareWomen)

[1] 4826.06

Python vs. R — The Results (using polars)

library(tidyverse)
library(rio)
DataTitanicR = import("Data/Titanic.csv") |>
  select(Sex, FareInPounds, Surv = Survived) |>
  mutate(FareInDollars = FareInPounds * 108.5) |>
  filter(Sex == "female")
MeanFareWomen = mean(DataTitanicR$FareInDollars)
print(MeanFareWomen)

[1] 4826.06

Note, polars is currently not supported in WASM

Was Chivalry Dead in 1912?

To answer the question, we develop a male and a female data frame and compare the survival rates.

In each data frame we would need only the variables Sex and Survived but we add also PasClass for additional analysis.

Exercise: The Male Data Frame

We select Sex, Survived, and PasClass=Pclass and filter for male:

The Female Data Frame

We select Sex, Survived, and PasClass=Pclass and filter for female:

Comparing the Survival Proportion of Males to Females

Hint: You could either calculate the female proportion as sum(DataFemale$Survived/nrow(DataFemale)) or mean(DataFemale$Survived).

Be critical with your own research

`PasClass` is a Confounder

The third class was deep in the hull of the Titanic with low survival chances and more men were traveling in that class. This makes PasClass a confounder. Therefore we have to analyze male and female survival by class: We have to filter for Sex and PasClass.

Survival Research for Passenger Class 1

Select Survived, Sex, PasClass=Pclass
Filter for PasClass and Sex (female and male)

Survival Research for Passenger Class 2

Select Survived, Sex, PasClass=Pclass
Filter for PasClass and Sex (female and male)

Survival Research for Passenger Class 3

Select Survived, Sex, PasClass=Pclass
Filter for PasClass and Sex (female and male)

Introduction to R and Positron

Learning Outcomes

Learning Outcomes

Learning Outcomes

Part 1: Install and Setup R and Positron

RStudio — Integrated Development Environment (IDE)

(does not have newest features of a modern IDE)

Positron — Integrated Development Environment (IDE)

Set Window Layout and Choose R Interpreter

RStudio Keybindings (needs to done only once)

Always Work with Folders

Use Command Palette to Create a New R or Quarto file

Open a File from an Existing Positron Folder

Positron: First Steps with an R File (.r)

An R file such as MyFile.r contains only R-code

Try Positron with a Quarto File (.qmd)

A Quarto file such as MyFile.qmd contains text as well as R-code

Try Positron with a Quarto File (.qmd)

WASM: R Runs in a Browser Including Libraries and Data

R Packages

Videos for the rio and the tidyverse Package

Using the rio and the tidyverse Package

Part 2: Data Types & Data Objects

Analogy: Data Types & Data Objects Example for Three Alcoholic Beverages

Data Types

Print

Calculate with Variables and Output with cat()

Exercise: cat() command and Single Value Objects with Different Data Types

Again: Data Types & Data Objects

Data Objects

Single Value Object

Vector-Objects

Data Frames (tibbles)

Data Frame from Titanic Data

Extracting the Vectors and Performing Calculations (numerical Vectors)

Extracting the Vectors and Performing Calculations (logical Vectors)

Summary Data Types

Summary Data Types and Objects

Part 3: The tidyverse and Piping

Basics of R Commands

Structure of R Commands

Use a Command with and without Argument Names

Getting Help about a Command (e.g., mean)

Important Commands from tidyverse/dplyr Package

Titanic Dataset

Example: Using the tidyverse for Data Analysis

The select() Command

The filter() Command

The mutate() Command

Summary

Piping Schema

Alternative: Piping

(will be used throughout the course/book)

Why R?

R and/or Python

Python vs. R — The task

Python vs. R — The Results (using pandas)

Python vs. R — The Results (using polars)

Was Chivalry Dead in 1912?

Exercise: The Male Data Frame

The Female Data Frame

Comparing the Survival Proportion of Males to Females

Be critical with your own research

PasClass is a Confounder

Survival Research for Passenger Class 1

Survival Research for Passenger Class 2

Survival Research for Passenger Class 3

Use Command Palette to Create a New `R` or `Quarto` file

Positron: First Steps with an R File `(.r)`

An R file such as `MyFile.r` contains only R-code

Try Positron with a Quarto File `(.qmd)`

A Quarto file such as `MyFile.qmd` contains text as well as R-code

Try Positron with a Quarto File `(.qmd)`

Videos for the `rio` and the `tidyverse` Package

Using the `rio` and the `tidyverse` Package

Calculate with Variables and Output with `cat()`

Exercise: `cat()` command and Single Value Objects with Different Data Types

Part 3: The `tidyverse` and Piping

Getting Help about a Command (e.g., `mean`)

Important Commands from `tidyverse`/`dplyr` Package

Example: Using the `tidyverse` for Data Analysis

The `select()` Command

The `filter()` Command

The `mutate()` Command

`PasClass` is a Confounder