Introduction to

Evangeline Warren

R Working Group

Jan 20th, 2023

1 / 43

Goals for this sessionLearn about...
2 / 43

Goals for this session

Learn about...
- basic R syntax
- different R objects (things that hold data) & indexing them
- useful functions for working with data

2 / 43

Goals for this session

Learn about...
- basic R syntax
- different R objects (things that hold data) & indexing them
- useful functions for working with data

Become familiar with R Studio & develop good coding habits

2 / 43

Goals for this session

Learn about...
- basic R syntax
- different R objects (things that hold data) & indexing them
- useful functions for working with data

Become familiar with R Studio & develop good coding habits
- R Studio is an additional program that provides many useful features for working with R
- (you need to download and install both R and R Studio)

2 / 43

R Studio3 / 43

R Studio

Let's dive in by starting R Studio and opening a new R script
- menu bar: File → New File → R Script
- (in R: File → New Script)

4 / 43

R Studio

Let's dive in by starting R Studio and opening a new R script
- menu bar: File → New File → R Script
- (in R: File → New Script)
You should now have 4 panes open (like on the next slide)
- Source -- Our script where we will type and save our comments & commands
- Console -- Where we can give R commands and where the output will appear
- Output -- File explorer, plots, help files, and more!
- Environments -- Useful information about the R session

4 / 43

downloaded from user guide on postit.co

5 / 43

R Studio: Good Habits

Add a comment to our new script:

# Comment: My R script from Working Group Session (1/20/2023)
# (R ignores all lines that begin with a pound/hash/number sign/#)

6 / 43

R Studio: Good Habits

Add a comment to our new script:

# Comment: My R script from Working Group Session (1/20/2023)
# (R ignores all lines that begin with a pound/hash/number sign/#)

Save our script
- menu bar: File → Save As...

6 / 43

R Studio: Good Habits

Add a comment to our new script:

# Comment: My R script from Working Group Session (1/20/2023)
# (R ignores all lines that begin with a pound/hash/number sign/#)

Save our script
- menu bar: File → Save As...
Set our working directory
- this is where R will start looking for & saving files (e.g., data files or plots)
- menu bar: Session → Set Working Directory →
  Choose Directory...

6 / 43

Basic R Syntax7 / 43

Basic R Syntax

R syntax takes the form

# object_name <- object_value  
mean_age <- 33

8 / 43

Basic R Syntax

R syntax takes the form

# object_name <- object_value  
mean_age <- 33

The symbol "<-" is called the assignment operator
- we are creating a new variable called mean_age and assigning it a value of 33
- mean_age = 33 will also work (but <- is the convention)

8 / 43

Basic R Syntax (cont.)

If we enter the name of a variable in the Console, then R will list the value(s)

> Mean_age2 <- 22  ## note: object names are case-sensitive
> Mean_age2

## [1] 22

9 / 43

Basic R Syntax (cont.)

If we enter the name of a variable in the Console, then R will list the value(s)

> Mean_age2 <- 22  ## note: object names are case-sensitive
> Mean_age2

## [1] 22

BUT we are in the business of good habits...

type this syntax into our script and (with the cursor on the same line) press the following keys together:
- On a Mac: <command> <enter>
- In Windows: <control> <enter> (in R Studio)
  <control> r (in the R app)
these keyboard shortcuts will run the syntax on the line in the Console
(or you can highlight a region)

9 / 43

Basic R Syntax: functions

We have seen a simple object for holding data, but R has many useful functions

ls()                         # list all the objects in memory
rm(Mean_age2)                # remove the object called Mean_age2
getwd()                      # print the working directory
dir()                        # list the files in the current directory
dir("../")                   # list the files in the parent directory
save.image("my_data.RData")  # save all the objects in memory
load("my_data.RData")        # load all the objects in the data file

10 / 43

Basic R Syntax: functions

We have seen a simple object for holding data, but R has many useful functions

ls()                         # list all the objects in memory
rm(Mean_age2)                # remove the object called Mean_age2
getwd()                      # print the working directory
dir()                        # list the files in the current directory
dir("../")                   # list the files in the parent directory
save.image("my_data.RData")  # save all the objects in memory
load("my_data.RData")        # load all the objects in the data file

Quick note:

suppose you create an object called abc that holds the value 2
then you load a file data.RData that also has an object named abc but holds the value 99
the first version of the object (abc holding 2) will get replaced

10 / 43

Basic R Syntax: help files

Google searches are a very effective way to find help
- and so is asking the R Working Group 😎

11 / 43

Basic R Syntax: help files

Google searches are a very effective way to find help
- and so is asking the R Working Group 😎
R documentation can be accessed in the Help tab in the Output pane

11 / 43

Basic R Syntax: help files

Google searches are a very effective way to find help
- and so is asking the R Working Group 😎
R documentation can be accessed in the Help tab in the Output pane
Some additional syntax and functions

?read.csv                     # show the help file for the function read.csv
help.search("weighted mean")  # search help files for the phrase'weighted mean'

11 / 43

Data Structures in R12 / 43

Data Structures: motivation

We are not going to solve the world's problems with a single number...

> all_ages <- c(22, 33, 44, 55)  # c() concatenates numbers together
> all_ages

## [1] 22 33 44 55

> mean(all_ages)                 # calculate the mean

## [1] 38.5

> all_ed <- c("HS", "Col", "Grad Sch", "HS")
> all_ed

## [1] "HS"       "Col"      "Grad Sch" "HS"

13 / 43

Data Structures: motivation (cont.)

R handles different types of data as well

> important_data <- c("OSU", "R", "Group", 4)
> important_data

## [1] "OSU"   "R"     "Group" "4"

Wait, what is going on here?

14 / 43

Data Structures: motivation (cont.)

R handles different types of data as well

> important_data <- c("OSU", "R", "Group", 4)
> important_data

## [1] "OSU"   "R"     "Group" "4"

Wait, what is going on here?

we are mixing different types of data & R assumes that we just forgot to wrap the 4 in quotation marks
sometimes R's assumptions are useful, sometimes they are not! 🤔

14 / 43

Data Structures: motivation (cont.)

Here is another example with missing data

> test_scores <- c(88, 99, 110, 66, NA)  # NA is for missing values
> mean_scores <- mean(test_scores)
> mean_scores / 100

## [1] NA

😾 Ugh! Why didn't R tell me there was a problem when I tried to calculate the mean?!?

15 / 43

Data Structures: motivation (cont.)

Here is another example with missing data

> test_scores <- c(88, 99, 110, 66, NA)  # NA is for missing values
> mean_scores <- mean(test_scores)
> mean_scores / 100

## [1] NA

😾 Ugh! Why didn't R tell me there was a problem when I tried to calculate the mean?!?

another R assumption
can you figure out how to calculate the mean for non-missing values? (help file is helpful 😄)

15 / 43

Data Structures: vectorsWe have been creating vectors when we use c() to concatenate data
16 / 43

Data Structures: vectors

We have been creating vectors when we use c() to concatenate data
Here are some more useful functions for working with vectors

> # test that we have a vector
> is.vector(test_scores)  # returns another data type: TRUE or FALSE (called logical)

## [1] TRUE

> summary(test_scores)    # numerical summary (less helpful for strings)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max.    NA's 
##   66.00   82.50   93.50   90.75  101.75  110.00       1

16 / 43

Data Structures: vectors (cont.)

> length(test_scores)     # how many elements in the vector

## [1] 5

> is.na(test_scores)      # test if each element is NA

## [1] FALSE FALSE FALSE FALSE  TRUE

> TRUE + TRUE + FALSE     # useful trick with logical objects (TRUE/FALSE)

## [1] 2

> n_missing <- sum(is.na(test_scores))
> n_missing

## [1] 1

17 / 43

Data Structures: indexing vectors

We can access the $i^{th}$ element in a vector with the syntax vector_name[ i ]

> test_scores[1]    # first element

## [1] 88

> test_scores[2]    # second element

## [1] 99

18 / 43

Data Structures: indexing vectors

We can access the $i^{th}$ element in a vector with the syntax vector_name[ i ]

> test_scores[1]    # first element

## [1] 88

> test_scores[2]    # second element

## [1] 99

> 1:3   # a vector of c(1, 2, 3)

## [1] 1 2 3

>       # so what will test_scores[3:1] give us?

18 / 43

Data Structures: indexing vectors (cont.)

The syntax 3:1 gives the vector c(3, 2, 1), so...

19 / 43

Data Structures: indexing vectors (cont.)

The syntax 3:1 gives the vector c(3, 2, 1), so...

> test_scores[3:1]  # returns 3rd element, then the 2nd, then the first

## [1] 110  99  88

> test_scores       # sanity check

## [1]  88  99 110  66  NA

19 / 43

Data Structures: indexing vectors (cont.)

The syntax 3:1 gives the vector c(3, 2, 1), so...

> test_scores[3:1]  # returns 3rd element, then the 2nd, then the first

## [1] 110  99  88

> test_scores       # sanity check

## [1]  88  99 110  66  NA

So what will the following command do? 🤔

test_scores[c(3, 5, 11)]

19 / 43

Data Structures: changing vectors

We can use indexing to change vectors as well, e.g., reassign the first element

> test_scores[1]  <- NA # change the first element to NA
> test_scores[1]

## [1] NA

20 / 43

Data Structures: changing vectors

We can use indexing to change vectors as well, e.g., reassign the first element

> test_scores[1]  <- NA # change the first element to NA
> test_scores[1]

## [1] NA

Again, we can use vectors to index as well:

index_missing_scores <- is.na(test_scores)  # create an index vector of TRUE & FALSE
test_scores[index_missing_scores] <- -99    # change NA to -99

20 / 43

Data Structures: changing vectors

We can use indexing to change vectors as well, e.g., reassign the first element

> test_scores[1]  <- NA # change the first element to NA
> test_scores[1]

## [1] NA

Again, we can use vectors to index as well:

index_missing_scores <- is.na(test_scores)  # create an index vector of TRUE & FALSE
test_scores[index_missing_scores] <- -99    # change NA to -99

Let's walk through this...
(🦉 but note a good habit would be to create a new vector, new_test_scores, so we can retain the original data!)

20 / 43

Data Structures: changing vectors (cont.)

> # create an index vector of TRUE & FALSE
> index_missing_scores <- is.na(test_scores)
> index_missing_scores

## [1]  TRUE FALSE FALSE FALSE  TRUE

21 / 43

Data Structures: changing vectors (cont.)

> # create an index vector of TRUE & FALSE
> index_missing_scores <- is.na(test_scores)
> index_missing_scores

## [1]  TRUE FALSE FALSE FALSE  TRUE

> # attach these 2 vectors together as columns
> cbind(index_missing_scores, test_scores)

##      index_missing_scores test_scores
## [1,]                    1          NA
## [2,]                    0          99
## [3,]                    0         110
## [4,]                    0          66
## [5,]                    1          NA

21 / 43

Data Structures: changing vectors (cont.)

> # create an index vector of TRUE & FALSE
> index_missing_scores <- is.na(test_scores)
> index_missing_scores

## [1]  TRUE FALSE FALSE FALSE  TRUE

> # attach these 2 vectors together as columns
> cbind(index_missing_scores, test_scores)

##      index_missing_scores test_scores
## [1,]                    1          NA
## [2,]                    0          99
## [3,]                    0         110
## [4,]                    0          66
## [5,]                    1          NA

with cbind we are actually creating a new data structure called a matrix

21 / 43

Data Structures: changing vectors (cont.)

> # create an index vector of TRUE & FALSE
> index_missing_scores <- is.na(test_scores)
> index_missing_scores

## [1]  TRUE FALSE FALSE FALSE  TRUE

> # attach these 2 vectors together as columns
> cbind(index_missing_scores, test_scores)

##      index_missing_scores test_scores
## [1,]                    1          NA
## [2,]                    0          99
## [3,]                    0         110
## [4,]                    0          66
## [5,]                    1          NA

with cbind we are actually creating a new data structure called a matrix
as we will see, matrices can only hold the same data type, so R changes TRUE/FALSE to 1/0 (respectively)

21 / 43

Data Structures: changing vectors (cont.)

> test_scores[index_missing_scores]  #  access all of the indices with TRUE

## [1] NA NA

22 / 43

Data Structures: changing vectors (cont.)

> test_scores[index_missing_scores]  #  access all of the indices with TRUE

## [1] NA NA

> # recode NA to -99
> test_scores[index_missing_scores] <- -99
> test_scores

## [1] -99  99 110  66 -99

22 / 43

Data Structures: changing vectors (recap)

When you want to change a vector, do the delta 2-step:

create an index vector that identifies the elements you want to change
- what data type should this vector hold?
- logical, i.e. TRUEs and FALSEs
assign new values to the vector using your vector of indices

23 / 43

Data Structures: vector recapWe are not going to become 💰 famous 💰 by working with
a single vector
24 / 43

Data Structures: vector recap

We are not going to become 💰 famous 💰 by working with a single vector
However, we have learned a powerful way to work with vectors, indexing, that extends to other types of data structures

24 / 43