03-tidyr

Author

Professor Shannon Ellis

Published

October 10, 2023

Tidy Data with `tidyr`

[ad] Data Science Student Society

Join DS3 at their Fall General Body Meeting to learn more about the events they’re offering this quarter, open board positions for the year, and free food! It will be happening on Wednesday (10/11) from 6-8pm, at PC Ballroom West

Q&A

Q: Is it possible to integrate to github using other systems than datahub? Datahub has already been spotty for me in this course and is notorious for slumping at critical pts in the quarter.
A: Yup. The same steps can be carried out by downloading RStudio onto your computer and connecting it with GitHub.

Q: Should we write in the console or in the rmd file first when writing code?
A: Great question! I’d suggest starting in the Rmd file and editing there. That way you don’t have to copy+paste once you get it right. It’s already there.

Q: How do you take notes for coding classes? I know there are lecture notes available, but how would you recommend taking notes for this class?
A: I would recommend opening a blank Rmd each day for class and saving it with the lecture number. I’d keep notes and things I tried in that file. But, I wouldn’t copy+paste everything, since the other lecture notes are available.

Course Announcements

Due Dates:

Lab 02 due Friday
HW01 now available; due Monday (10/16; 11:59 PM)
Lecture Participation survey open until Thursday

. . .

Notes:

Lab01 scores and feedback posted
Datahub: Launch RStudio (possible solution?)
Staff office hours updated (see Canvas or website)

Student Comment

I have been struggling to grasp the material in the course. It feels like we are diving into the content in the labs, but I don’t even feel like I truly understand what I’m doing. It often seems like I’m just copying and pasting code from the website without a clear understanding of the bigger picture. I’m particularly stuck because I feel like I don’t have a solid grasp of the fundamental concepts of coding in R; it feels so new. I understand that the pace of the course may be challenging, but I think a bit more stronger focus on the foundational aspects of coding in R would greatly benefit students like me who are struggling with the content. I’m looking forward to the course and I hope I can grasp the content as we go through the next week. I’m concerned about learning the material and also how that may affect my grade.

. . .

Let’s see how y’all feel in a week. The first week can be a lot in this course. Often, students feel a lot more comfortable come week 3.

Student Survey

89% know Python; 15% know R; most (but not all!) have programmed before
64% feel confident about effective data science communication
Reasons for taking course: learn R, add to resume, analyze data, improve data science skills

. . .

My favorite boring facts:

I was actually born on my birthday
i don’t like to eat eggs but my roommate loves them
I like to have a midday nap.
I can raise my eyebrows really well
I eat peanut butter straight from the jar
i have a jack russell terrier.. named jack (we weren’t feeling creative)

Tidy Data

The opinionated tidyverse is named as such b/c it assumes/necessitates your data be “tidy”.

. . .

Tidy datasets are all alike, but every messy dataset is messy in its own way. —- Hadley Wickham

. . .

Each variable must have its own column.
Each observation must have its own row.
Each value must have its own cell.

Source: https://r4ds.had.co.nz/tidy-data.html

Tidy or not?

❓ Given the rules discussed, is the cat_lovers dataset tidy?

cat_lovers <- read_csv("https://raw.githubusercontent.com/COGS137/datasets/main/cat-lovers.csv")

Rows: 60 Columns: 3
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
chr (3): name, number_of_cats, handedness

ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

cat_lovers |> datatable()

. . .

❓ Given the rules discussed, is the bike dataset tidy?

bike <- read_csv2("https://raw.githubusercontent.com/COGS137/datasets/main/nc_bike_crash.csv", 
                  na = c("NA", "", "."))

ℹ Using "','" as decimal and "'.'" as grouping mark. Use `read_delim()` for more control.

Rows: 5716 Columns: 54
── Column specification ────────────────────────────────────────────────────────
Delimiter: ";"
chr  (44): AmbulanceR, BikeAge_Gr, Bike_Alc_D, Bike_Dir, Bike_Injur, Bike_Po...
dbl   (8): FID, OBJECTID, Bike_Age, Crash_Hour, Crash_Ty_1, Crash_Year, Drvr...
dttm  (1): Crash_Time
date  (1): Crash_Date

ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

bike |> datatable()

Warning in instance$preRenderHook(instance): It seems your data is too big for
client-side DataTables. You may consider server-side processing:
https://rstudio.github.io/DT/server.html

Summary tables

❓ Which is a dataset? Which is a summary table?

Your Turn

There are four representations of the same data/information provided in the tidyr packages: table1, table2, table3, and the combination of table4a and table4b. Given what we’ve discussed, which is the best (tidiest) way to represent these data?

Put a green sticky on the front of your computer when you’re done. Put a pink if you want help/have a question.

Common issues

One variable might be spread across multiple columns.
One observation might be scattered across multiple rows.

. . .

Solution: pivoting!

Pivoting

For when some of the column names are not names of variables, but values of a variable…

table4a

# A tibble: 3 × 3
  country     `1999` `2000`
  <chr>        <dbl>  <dbl>
1 Afghanistan    745   2666
2 Brazil       37737  80488
3 China       212258 213766

table4a |> 
  pivot_longer(c(`1999`, `2000`), names_to = "year", values_to = "cases")

# A tibble: 6 × 3
  country     year   cases
  <chr>       <chr>  <dbl>
1 Afghanistan 1999     745
2 Afghanistan 2000    2666
3 Brazil      1999   37737
4 Brazil      2000   80488
5 China       1999  212258
6 China       2000  213766

❓ Why are there backticks around the years? (Note: we have not discussed this yet)

For when an observation is scattered across multiple rows…

table2

# A tibble: 12 × 4
   country      year type            count
   <chr>       <dbl> <chr>           <dbl>
 1 Afghanistan  1999 cases             745
 2 Afghanistan  1999 population   19987071
 3 Afghanistan  2000 cases            2666
 4 Afghanistan  2000 population   20595360
 5 Brazil       1999 cases           37737
 6 Brazil       1999 population  172006362
 7 Brazil       2000 cases           80488
 8 Brazil       2000 population  174504898
 9 China        1999 cases          212258
10 China        1999 population 1272915272
11 China        2000 cases          213766
12 China        2000 population 1280428583

table2 |>
    pivot_wider(names_from = type, values_from = count)

# A tibble: 6 × 4
  country      year  cases population
  <chr>       <dbl>  <dbl>      <dbl>
1 Afghanistan  1999    745   19987071
2 Afghanistan  2000   2666   20595360
3 Brazil       1999  37737  172006362
4 Brazil       2000  80488  174504898
5 China        1999 212258 1272915272
6 China        2000 213766 1280428583

❓ Why aren’t there quotes around column names here…but there were in pivot_longer? (Note: we have not discussed this yet.)

wide data contains values that do not repeat in the first column.
long format contains values that do repeat in the first column.

Both are good/helpful! We’ll return to this idea and discuss more during dataviz next week.

Briefly:

wide data: analysis
long data: plotting

For when multiple pieces of information are stored in a single column…

table3

# A tibble: 6 × 3
  country      year rate             
  <chr>       <dbl> <chr>            
1 Afghanistan  1999 745/19987071     
2 Afghanistan  2000 2666/20595360    
3 Brazil       1999 37737/172006362  
4 Brazil       2000 80488/174504898  
5 China        1999 212258/1272915272
6 China        2000 213766/1280428583

table3 |> 
  separate(rate, into = c("cases", "population"))

# A tibble: 6 × 4
  country      year cases  population
  <chr>       <dbl> <chr>  <chr>     
1 Afghanistan  1999 745    19987071  
2 Afghanistan  2000 2666   20595360  
3 Brazil       1999 37737  172006362 
4 Brazil       2000 80488  174504898 
5 China        1999 212258 1272915272
6 China        2000 213766 1280428583

…but…but…cases and population should be numeric…

table3 |> 
  separate(rate, into = c("cases", "population"), convert = TRUE)

# A tibble: 6 × 4
  country      year  cases population
  <chr>       <dbl>  <int>      <int>
1 Afghanistan  1999    745   19987071
2 Afghanistan  2000   2666   20595360
3 Brazil       1999  37737  172006362
4 Brazil       2000  80488  174504898
5 China        1999 212258 1272915272
6 China        2000 213766 1280428583

Unite is the opposite…it combines data stored across multiple columns.

The general syntax is:

df |>
  unite(new_col, first_col, second_col)

Joins

If we look at table4a, it’s missing the population information. That’s stored in a separate table…table4b

table4b

# A tibble: 3 × 3
  country         `1999`     `2000`
  <chr>            <dbl>      <dbl>
1 Afghanistan   19987071   20595360
2 Brazil       172006362  174504898
3 China       1272915272 1280428583

…which is also in the “wide” format

. . .

…so we pivot both tables longer

tidy4a <- table4a |> 
  pivot_longer(c(`1999`, `2000`), names_to = "year", values_to = "cases")

tidy4b <- table4b |> 
  pivot_longer(c(`1999`, `2000`), names_to = "year", values_to = "population")

tidy4b

# A tibble: 6 × 3
  country     year  population
  <chr>       <chr>      <dbl>
1 Afghanistan 1999    19987071
2 Afghanistan 2000    20595360
3 Brazil      1999   172006362
4 Brazil      2000   174504898
5 China       1999  1272915272
6 China       2000  1280428583

. . .

…but how do we get them into a single tidy dataset?

. . .

A join!

left_join(tidy4a, tidy4b)

Joining with `by = join_by(country, year)`

# A tibble: 6 × 4
  country     year   cases population
  <chr>       <chr>  <dbl>      <dbl>
1 Afghanistan 1999     745   19987071
2 Afghanistan 2000    2666   20595360
3 Brazil      1999   37737  172006362
4 Brazil      2000   80488  174504898
5 China       1999  212258 1272915272
6 China       2000  213766 1280428583

Source: R4DS

The Data: `nycflights13`

library(nycflights13)

airlines : links airline to two letter code
airports : ID’ed by FAA code
planes : ID’ed by tailnum
airport : weather each hour; id’ed by two letter airport code

. . .

Image Source: https://r4ds.had.co.nz/relational-data.html)

. . .

flights connects to planes via a single variable, tailnum.
flights connects to airlines through the carrier variable.
flights connects to airports in two ways: via the origin and dest variables.
flights connects to weather via origin (the location), and year, month, day and hour (the time).

Mutating Joins

mutating joins - add new variables to a data frame from matching observations in another

. . .

For simplicity, we’ll work with only a handful of columns…

flights |> 
  select(year:day, hour, tailnum, carrier) |> 
  left_join(airlines, by = "carrier")

# A tibble: 336,776 × 7
    year month   day  hour tailnum carrier name                    
   <int> <int> <int> <dbl> <chr>   <chr>   <chr>                   
 1  2013     1     1     5 N14228  UA      United Air Lines Inc.   
 2  2013     1     1     5 N24211  UA      United Air Lines Inc.   
 3  2013     1     1     5 N619AA  AA      American Airlines Inc.  
 4  2013     1     1     5 N804JB  B6      JetBlue Airways         
 5  2013     1     1     6 N668DN  DL      Delta Air Lines Inc.    
 6  2013     1     1     5 N39463  UA      United Air Lines Inc.   
 7  2013     1     1     6 N516JB  B6      JetBlue Airways         
 8  2013     1     1     6 N829AS  EV      ExpressJet Airlines Inc.
 9  2013     1     1     6 N593JB  B6      JetBlue Airways         
10  2013     1     1     6 N3ALAA  AA      American Airlines Inc.  
# ℹ 336,766 more rows

There is now a new column name…coming from the airlines data frame.

. . .

left_join:

keeps all rows in first df (here: flights)
adds all matching information from second df (here: airlines); adds NAs for any observations not in airlines

. . .

Other joins:

right_join: keeps all observations in second df full_join: keeps all observations in either df

. . .

inner_join:

takes only rows in both dfs

Recap

Do you understand what constitutes tidy data?
Can you identify what needs to be done to take a dataset from untidy to tidy?
What is the difference been long data and wide data?
Can I take long data to wide data? And wide to long?
Can I carry out mutating joins on data?

Tidy Data with tidyr

[ad] Data Science Student Society

Q&A

Course Announcements

Student Comment

Student Survey

Suggested Reading

Tidy Data

Tidy or not?

Summary tables

Your Turn

Common issues

Pivoting

Separating & Uniting

Joins

The Data: nycflights13

Mutating Joins

Recap

Tidy Data with `tidyr`

The Data: `nycflights13`