CONJ620: CM 1.2Introduction to the TidyverseAlison Hill1 / 37

I'm assuming...

You have working versions of:

R and
RStudio

installed on your computer.

2 / 37

Today's in-class lab

Our notepad: http://bit.ly/conj620-cm012

3 / 37

Today's in-class lab

Our notepad: http://bit.ly/conj620-cm012
Graphics and statistics for cardiology: comparing categorical and continuous variables
- Full text
- doi:10.1136/heartjnl-2015-308104
- Site: http://faculty.washington.edu/kenrice/heartgraphs/

3 / 37

Today's in-class lab

Our notepad: http://bit.ly/conj620-cm012
Graphics and statistics for cardiology: comparing categorical and continuous variables
- Full text
- doi:10.1136/heartjnl-2015-308104
- Site: http://faculty.washington.edu/kenrice/heartgraphs/

The data is from NHANES: the CDC's National Health and Nutrition Examination Survey

3 / 37

Set-up

I provide this code in your lab worksheet

library(tidyverse)
heart <- read_csv("http://faculty.washington.edu/kenrice/heartgraphs/nhanesmedium.csv", 
                  na = ".")

Some notes:

I'm going to ask that you trust me with this worksheet! You haven't learned about this kind of document yet, called an R Markdown (.Rmd) file- please just go with it! I promise we'll actually explain it in a later lab 😇
Don't forget the notepad!: http://bit.ly/conj620-cm012

4 / 37

Data Dictionary

From the data dictionary:

BPXSAR: systolic blood pressure (mmHg)
BPXDAR: diastolic blood pressure (mmHg)
BPXDI1, BPXDI2: two diastolic blood pressure readings
race_ethc: race/ethnicity, coded as:
- Hispanic,
- White non-Hispanic,
- Black non-Hispanic and
- Other
gender: sex, coded as Male/Female
DR1TFOLA: folate intake (μg/day)
RIAGENDR: sex, coded as 1/2
BMXBMI: body mass index (kg/m2)
RIDAGEY: age (years)

5 / 37

Chapter 1: Data Wrangling

print a tibble
- heart
install a package
- install.packages("dplyr")
- do 1x per machine
load an installed package
- library(dplyr)
- do 1x per work session
assign a variable a name (<-)

6 / 37

Chapter 1: Data Wrangling

print a tibble
- heart
install a package
- install.packages("dplyr")
- do 1x per machine
load an installed package
- library(dplyr)
- do 1x per work session
assign a variable a name (<-)

dplyr::filter
dplyr::arrange
dplyr::mutate

6 / 37

Plus: `%>%`

image courtesy @LegoRLady

7 / 37

`%>%`

The pipe

"dataframe first, dataframe once"

8 / 37

`%>%`

The pipe

"dataframe first, dataframe once"

library(dplyr)

8 / 37

`%>%`

The pipe

"dataframe first, dataframe once"

library(dplyr)

RStudio Keyboard Shortcuts:

OSX: CMD + SHIFT + M

Else: CTRL + SHIFT + M

8 / 37

Nesting a dataframe inside a function is hard to read.

slice(heart, 1)

# A tibble: 1 x 10
  BPXSAR BPXDAR BPXDI1 BPXDI2 race_ethc    gender DR1TFOLA RIAGENDR BMXBMI
   <dbl>  <dbl>  <dbl>  <dbl> <chr>        <chr>     <dbl>    <dbl>  <dbl>
1   129.   50.7    48.    48. White non-H… Male       334.       1.   19.7
# ... with 1 more variable: RIDAGEYR <dbl>

9 / 37

Nesting a dataframe inside a function is hard to read.

slice(heart, 1)

# A tibble: 1 x 10
  BPXSAR BPXDAR BPXDI1 BPXDI2 race_ethc    gender DR1TFOLA RIAGENDR BMXBMI
   <dbl>  <dbl>  <dbl>  <dbl> <chr>        <chr>     <dbl>    <dbl>  <dbl>
1   129.   50.7    48.    48. White non-H… Male       334.       1.   19.7
# ... with 1 more variable: RIDAGEYR <dbl>

Here, the "sentence" starts with a verb.

9 / 37

Nesting a dataframe inside a function is hard to read.

slice(heart, 1)

# A tibble: 1 x 10
  BPXSAR BPXDAR BPXDI1 BPXDI2 race_ethc    gender DR1TFOLA RIAGENDR BMXBMI
   <dbl>  <dbl>  <dbl>  <dbl> <chr>        <chr>     <dbl>    <dbl>  <dbl>
1   129.   50.7    48.    48. White non-H… Male       334.       1.   19.7
# ... with 1 more variable: RIDAGEYR <dbl>

Here, the "sentence" starts with a verb.

Piping a dataframe into a function lets you read L to R

heart %>% slice(1)

# A tibble: 1 x 10
  BPXSAR BPXDAR BPXDI1 BPXDI2 race_ethc    gender DR1TFOLA RIAGENDR BMXBMI
   <dbl>  <dbl>  <dbl>  <dbl> <chr>        <chr>     <dbl>    <dbl>  <dbl>
1   129.   50.7    48.    48. White non-H… Male       334.       1.   19.7
# ... with 1 more variable: RIDAGEYR <dbl>

9 / 37

Nesting a dataframe inside a function is hard to read.

slice(heart, 1)

# A tibble: 1 x 10
  BPXSAR BPXDAR BPXDI1 BPXDI2 race_ethc    gender DR1TFOLA RIAGENDR BMXBMI
   <dbl>  <dbl>  <dbl>  <dbl> <chr>        <chr>     <dbl>    <dbl>  <dbl>
1   129.   50.7    48.    48. White non-H… Male       334.       1.   19.7
# ... with 1 more variable: RIDAGEYR <dbl>

Here, the "sentence" starts with a verb.

Piping a dataframe into a function lets you read L to R

heart %>% slice(1)

# A tibble: 1 x 10
  BPXSAR BPXDAR BPXDI1 BPXDI2 race_ethc    gender DR1TFOLA RIAGENDR BMXBMI
   <dbl>  <dbl>  <dbl>  <dbl> <chr>        <chr>     <dbl>    <dbl>  <dbl>
1   129.   50.7    48.    48. White non-H… Male       334.       1.   19.7
# ... with 1 more variable: RIDAGEYR <dbl>

Now, the "sentence" starts with a noun.

9 / 37

Sequences of functions make you read inside out

slice(filter(heart, gender == "Male"), 1)

# A tibble: 1 x 10
  BPXSAR BPXDAR BPXDI1 BPXDI2 race_ethc    gender DR1TFOLA RIAGENDR BMXBMI
   <dbl>  <dbl>  <dbl>  <dbl> <chr>        <chr>     <dbl>    <dbl>  <dbl>
1   129.   50.7    48.    48. White non-H… Male       334.       1.   19.7
# ... with 1 more variable: RIDAGEYR <dbl>

10 / 37

Sequences of functions make you read inside out

slice(filter(heart, gender == "Male"), 1)

# A tibble: 1 x 10
  BPXSAR BPXDAR BPXDI1 BPXDI2 race_ethc    gender DR1TFOLA RIAGENDR BMXBMI
   <dbl>  <dbl>  <dbl>  <dbl> <chr>        <chr>     <dbl>    <dbl>  <dbl>
1   129.   50.7    48.    48. White non-H… Male       334.       1.   19.7
# ... with 1 more variable: RIDAGEYR <dbl>

Chaining functions together lets you read L to R

heart %>% filter(gender == "Male") %>% slice(1)

# A tibble: 1 x 10
  BPXSAR BPXDAR BPXDI1 BPXDI2 race_ethc    gender DR1TFOLA RIAGENDR BMXBMI
   <dbl>  <dbl>  <dbl>  <dbl> <chr>        <chr>     <dbl>    <dbl>  <dbl>
1   129.   50.7    48.    48. White non-H… Male       334.       1.   19.7
# ... with 1 more variable: RIDAGEYR <dbl>

10 / 37

"dataframe first, dataframe once"

11 / 37

heart %>% filter(gender == "Male") %>% slice(1)

# A tibble: 1 x 10
  BPXSAR BPXDAR BPXDI1 BPXDI2 race_ethc    gender DR1TFOLA RIAGENDR BMXBMI
   <dbl>  <dbl>  <dbl>  <dbl> <chr>        <chr>     <dbl>    <dbl>  <dbl>
1   129.   50.7    48.    48. White non-H… Male       334.       1.   19.7
# ... with 1 more variable: RIDAGEYR <dbl>

12 / 37

heart %>% filter(gender == "Male") %>% slice(1)

# A tibble: 1 x 10
  BPXSAR BPXDAR BPXDI1 BPXDI2 race_ethc    gender DR1TFOLA RIAGENDR BMXBMI
   <dbl>  <dbl>  <dbl>  <dbl> <chr>        <chr>     <dbl>    <dbl>  <dbl>
1   129.   50.7    48.    48. White non-H… Male       334.       1.   19.7
# ... with 1 more variable: RIDAGEYR <dbl>

This does the same thing:

heart %>% filter(.data = ., gender == "Male") %>% slice(.data = ., 1)

12 / 37

heart %>% filter(gender == "Male") %>% slice(1)

# A tibble: 1 x 10
  BPXSAR BPXDAR BPXDI1 BPXDI2 race_ethc    gender DR1TFOLA RIAGENDR BMXBMI
   <dbl>  <dbl>  <dbl>  <dbl> <chr>        <chr>     <dbl>    <dbl>  <dbl>
1   129.   50.7    48.    48. White non-H… Male       334.       1.   19.7
# ... with 1 more variable: RIDAGEYR <dbl>

This does the same thing:

heart %>% filter(.data = ., gender == "Male") %>% slice(.data = ., 1)

So does this:

heart %>% filter(., gender == "Male") %>% slice(., 1)

12 / 37

attach()

heart$gender or other variants

13 / 37

⌛️Let's review some helpful functions for filter14 / 37

Base R + Tidyverse15 / 37

💡First:Logical Operators16 / 37

?base::Logic

Operator	Description	Usage
&	and	x & y
\|	or	x \| y
xor	exactly x or y	xor(x, y)
!	not	!x

17 / 37

Logical or (|) is inclusive, so x | y really means:

x or
y or
both x & y

Exclusive or (xor) is exclusive, so xor(x, y) really means:

x or
y...
but not both x & y

x <- c(0, 1, 0, 1)
y <- c(0, 0, 1, 1)
boolean_or <- x | y
exclusive_or <- xor(x, y)
cbind(x, y, boolean_or, exclusive_or)

     x y boolean_or exclusive_or
[1,] 0 0          0            0
[2,] 1 0          1            1
[3,] 0 1          1            1
[4,] 1 1          1            0

18 / 37

💡Second:Comparisons19 / 37

?Comparison

Operator	Description	Usage
<	less than	x < y
<=	less than or equal to	x <= y
>	greater than	x > y
>=	greater than or equal to	x >= y
==	exactly equal to	x == y
!=	not equal to	x != y
%in%	group membership*	x %in% y
is.na	is missing	is.na(x)
!is.na	is not missing	!is.na(x)

*(shortcut to using | repeatedly with ==)

20 / 37

⌛️Let's review mutate21 / 37

3 ways to `mutate`

Create a new variable with a specific value
Create a new variable based on other variables
Change an existing variable

22 / 37

3 ways to `mutate`

Create a new variable with a specific value
Create a new variable based on other variables
Change an existing variable

heart_bp <- heart %>% 
  select(BPXDI1, BPXDI2)
heart_bp %>% 
  mutate(year = 2015)

# A tibble: 200 x 3
   BPXDI1 BPXDI2  year
    <dbl>  <dbl> <dbl>
 1    48.    48. 2015.
 2    76.    78. 2015.
 3    76.    76. 2015.
 4    64.    56. 2015.
 5    54.    56. 2015.
 6    80.    78. 2015.
 7    52.    NA  2015.
 8    NA     80. 2015.
 9    76.    NA  2015.
10    90.    80. 2015.
# ... with 190 more rows

22 / 37

3 ways to `mutate`

Create a new variable with a specific value
Create a new variable based on other variables
Change an existing variable

23 / 37

3 ways to `mutate`

Create a new variable with a specific value
Create a new variable based on other variables
Change an existing variable

heart_bp %>% 
  mutate(bp_ratio = BPXDI1 / BPXDI2)

# A tibble: 200 x 3
   BPXDI1 BPXDI2 bp_ratio
    <dbl>  <dbl>    <dbl>
 1    48.    48.    1.00 
 2    76.    78.    0.974
 3    76.    76.    1.00 
 4    64.    56.    1.14 
 5    54.    56.    0.964
 6    80.    78.    1.03 
 7    52.    NA    NA    
 8    NA     80.   NA    
 9    76.    NA    NA    
10    90.    80.    1.12 
# ... with 190 more rows

23 / 37

3 ways to `mutate`

Create a new variable with a specific value
Create a new variable based on other variables
Change an existing variable

24 / 37

3 ways to `mutate`

Create a new variable with a specific value
Create a new variable based on other variables
Change an existing variable

heart_bp %>% 
  mutate(bp_ratio = bp_ratio * 100)

# A tibble: 200 x 3
   BPXDI1 BPXDI2 bp_ratio
    <dbl>  <dbl>    <dbl>
 1    48.    48.    100. 
 2    76.    78.     97.4
 3    76.    76.    100. 
 4    64.    56.    114. 
 5    54.    56.     96.4
 6    80.    78.    103. 
 7    52.    NA      NA  
 8    NA     80.     NA  
 9    76.    NA      NA  
10    90.    80.    112. 
# ... with 190 more rows

24 / 37

To make your `mutate` "stick"

heart_bp <- heart_bp %>% 
  mutate(bp_ratio = BPXDI1 / BPXDI2)

25 / 37

⌛️Let's review some helpful functions for mutate26 / 37

Remember:Base R + Tidyverse27 / 37

💡

First:

Arithmetic

especially useful for mutate

See:

http://r4ds.had.co.nz/transform.html#mutate-funs

28 / 37

?Arithmetic

Operator	Description	Usage
+	addition	x + y
-	subtraction	x - y
*	multiplication	x * y
/	division	x / y
^	raised to the power of	x ^ y
abs	absolute value	abs(x)
%/%	integer division	x %/% y
%%	remainder after division	x %% y

5 %/% 2 # 2 goes into 5 two times with...

[1] 2

5 %% 2 # 1 left over

[1] 1

29 / 37

Chapter 2: Data Visualization

all ggplot2

aes(x = , y = ) (aesthetics)
aes(x = , y = , color = ) (add color)
aes(x = , y = , size = ) (add size)
+ facet_wrap(~ ) (facetting)

30 / 37

Old School ¹

Sketch the graphics below on paper, where the x-axis is variable age_yrs and the y-axis is variable systolic_bp

# A tibble: 4 x 4
  age_yrs systolic_bp bmi_z gender
    <dbl>       <dbl> <dbl> <chr> 
1      8.         80.    1. male  
2      9.         90.    2. male  
3     10.        100.    3. female
4     11.        110.    4. female

A scatter plot
A scatter plot where the color of the points corresponds to gender
A scatter plot where the size of the points corresponds to bmi_z
A scatter plot facetted by gender

[1] Shamelessly borrowed with much appreciation to Chester Ismay

31 / 37

1. A scatterplot

library(ggplot2)
ggplot(nhanes, aes(age_yrs, systolic_bp)) + 
  geom_point()

32 / 37

1. A scatterplot

library(ggplot2)
ggplot(nhanes, aes(age_yrs, systolic_bp)) + 
  geom_point()

32 / 37

2. `color` points by `gender`

library(ggplot2)
ggplot(nhanes, aes(age_yrs, systolic_bp, color = gender)) + 
  geom_point()

33 / 37

2. `color` points by `gender`

library(ggplot2)
ggplot(nhanes, aes(age_yrs, systolic_bp, color = gender)) + 
  geom_point()

33 / 37

3. `size` points by `bmi_z`

library(ggplot2)
ggplot(nhanes, aes(age_yrs, systolic_bp, size = bmi_z)) +  
  geom_point()

34 / 37

3. `size` points by `bmi_z`

library(ggplot2)
ggplot(nhanes, aes(age_yrs, systolic_bp, size = bmi_z)) +  
  geom_point()

34 / 37

4. `facet_wrap` by `gender`

library(ggplot2)
ggplot(nhanes, aes(age_yrs, systolic_bp)) +  
  geom_point() +
  facet_wrap(~gender)

35 / 37

4. `facet_wrap` by `gender`

library(ggplot2)
ggplot(nhanes, aes(age_yrs, systolic_bp)) +  
  geom_point() +
  facet_wrap(~gender)

35 / 37

Work your way through the lab worksheet now

View it here

Download the file on Sakai to work on locally

36 / 37

🔨

More Resources

37 / 37

Help

Keyboard shortcuts

↑, ←, Pg Up, k

Go to previous slide

↓, →, Pg Dn, Space, j

Go to next slide

Home

Go to first slide

End

Go to last slide

Number + Return

Go to specific slide

b / m / f

Toggle blackout / mirrored / fullscreen mode

Clone slideshow

Toggle presenter mode

Restart the presentation timer

?, h

Toggle this help

CONJ620: CM 1.2

Introduction to the Tidyverse

Alison Hill

I'm assuming...

Today's in-class lab

Today's in-class lab

Today's in-class lab

Set-up

Data Dictionary

Chapter 1: Data Wrangling

Chapter 1: Data Wrangling

Plus: %>%

%>%

The pipe

%>%

The pipe

%>%

The pipe

"dataframe first, dataframe once"

⌛️

Let's review some helpful functions for filter

Base R + Tidyverse

💡

First:

Logical Operators

💡

Second:

Comparisons

⌛️

Let's review mutate

3 ways to mutate

3 ways to mutate

3 ways to mutate

3 ways to mutate

3 ways to mutate

3 ways to mutate

To make your mutate "stick"

⌛️

Let's review some helpful functions for mutate

Remember:

Base R + Tidyverse

💡

First:

Arithmetic

Chapter 2: Data Visualization

Old School 1

1. A scatterplot

1. A scatterplot

2. color points by gender

2. color points by gender

3. size points by bmi_z

3. size points by bmi_z

4. facet_wrap by gender

4. facet_wrap by gender

Work your way through the lab worksheet now

🔨

More Resources

I'm assuming...

Help

Plus: `%>%`

`%>%`

`%>%`

`%>%`

Let's review some helpful functions for `filter`

Let's review `mutate`

3 ways to `mutate`

3 ways to `mutate`

3 ways to `mutate`

3 ways to `mutate`

3 ways to `mutate`

3 ways to `mutate`

To make your `mutate` "stick"

Let's review some helpful functions for `mutate`

Old School ¹

2. `color` points by `gender`

2. `color` points by `gender`

3. `size` points by `bmi_z`

3. `size` points by `bmi_z`

4. `facet_wrap` by `gender`

4. `facet_wrap` by `gender`