Objectives

Learn essential data manipulation functions of dplyr:
- select() columns
- filter() rows
- arrange() rows
- mutate() columns
- group_by() rows
Learn about the pipe operator (%>%) and how to use it to chain together multiple functions.

Loading data

In the previous lesson we used the read_csv() function to load the gm97 data. Let’s do that again:

# =========================================================================
# Data Transformation with dplyr
# =========================================================================

# -------------------------------------------------------------------------
# Ensure libraries and data are loaded

library(tidyverse)
gm97 = read_csv('data/gapminder_1997.csv')

Rows: 142 Columns: 6
── Column specification ─────────────────────────────────────────────────────────────────────────────────
Delimiter: ","
chr (2): country, continent
dbl (4): year, pop, lifeExp, gdpPercap

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.

This time, let’s look more closely at the data. If we type the name of the object and evaluate it, we’ll see a preview:

# -------------------------------------------------------------------------
# Preview the data

gm97

# A tibble: 142 × 6
   country      year       pop continent lifeExp gdpPercap
   <chr>       <dbl>     <dbl> <chr>       <dbl>     <dbl>
 1 Afghanistan  1997  22227415 Asia         41.8      635.
 2 Albania      1997   3428038 Europe       73.0     3193.
 3 Algeria      1997  29072015 Africa       69.2     4797.
 4 Angola       1997   9875024 Africa       41.0     2277.
 5 Argentina    1997  36203463 Americas     73.3    10967.
 6 Australia    1997  18565243 Oceania      78.8    26998.
 7 Austria      1997   8069876 Europe       77.5    29096.
 8 Bahrain      1997    598561 Asia         73.9    20292.
 9 Bangladesh   1997 123315288 Asia         59.4      973.
10 Belgium      1997  10199787 Europe       77.5    27561.
# ℹ 132 more rows

Checkpoint

The output shows we have “a tibble” and its dimensions are 142 rows by 6 columns. We then see:

Column headings (country, year, etc)
Column data types (chr, dbl, etc)
The entries of the first 10 rows.

Data in the form of a table is very common in R. The “base R” type is called a data.frame. The tidyverse extends on this notion with a tibble. The commands we’ll learn in this lesson work on data.frames and tibbles. We’ll use data.frame, tibble, and “data table” interchangeably throughout the lessons.

Exploring data tables

After reading in data, it’s a good habit to preview it, look at summaries of it, and, in so doing, look for issues. When data contains hundreds of rows, it’s a good idea to get a sense for the values and ranges of the data.

Get stats with `summary()`

The summary() function, from base R, takes a data table as input and will summarize the columns automatically depending on their data type.

# -------------------------------------------------------------------------
# Summarize all column values

summary(gm97)

   country               year           pop             continent        
 Length:142         Min.   :1997   Min.   :1.456e+05   Length:142        
 Class :character   1st Qu.:1997   1st Qu.:3.770e+06   Class :character  
 Mode  :character   Median :1997   Median :9.735e+06   Mode  :character  
                    Mean   :1997   Mean   :3.884e+07                     
                    3rd Qu.:1997   3rd Qu.:2.431e+07                     
                    Max.   :1997   Max.   :1.230e+09                     
    lifeExp        gdpPercap      
 Min.   :36.09   Min.   :  312.2  
 1st Qu.:55.63   1st Qu.: 1366.8  
 Median :69.39   Median : 4781.8  
 Mean   :65.01   Mean   : 9090.2  
 3rd Qu.:74.17   3rd Qu.:12022.9  
 Max.   :80.69   Max.   :41283.2

We see that for the numeric data we get quantile information and the mean, whereas for character data we simply get the length of the column. For data with NAs, the number thereof would be displayed.

Get stats with `summarize()`

If we wanted to know the mean life expectancy in the dataset, we could use the the dplyr summarize() function, specifying that we want the mean():

# -------------------------------------------------------------------------
# Calculate the mean life expectancy over the dataset

summarize(gm97, avgLifeExp = mean(lifeExp))

# A tibble: 1 × 1
  avgLifeExp
       <dbl>
1       65.0

We can use any column name as the input to mean(), so long as it’s a numeric, and we get a new tibble with our value returned as a column, named as specified. Note, this matches the result of summary() with some difference in the number of significant digits displayed. Alternatively, we could have simply applied mean() directly to the column lifeExp:

# -------------------------------------------------------------------------
# Base R approach to calculating the mean life expectancy

mean(gm97$lifeExp)

[1] 65.01468

This is the base R way, with the $ being how columns of a data table are accessed. The mean() function returns a number, whereas summarize() returns a tibble. In the tidyverse, a function whose input is a tibble will also have a tibble as output.

Checkpoint

Tip: Reference columns by name

While it is possible to access columns by their numerical index, this is less prefereable than using the explicit name of the column because altering the table could change the column indices. Depending on the situation, we might not even get an error, and instead would be simply operating on the wrong column. A dangerous situation to be in.

Find unique values with `distinct()`

Data is often subject to errors. To quickly catch data entry problems, the distinct() function can be used to show all unique values in the column of a table. Let’s take a look at the distinct values of the continents column:

# -------------------------------------------------------------------------
# Determine the distinct continents present in the dataset

distinct(gm97, continent)

# A tibble: 5 × 1
  continent
  <chr>    
1 Asia     
2 Europe   
3 Africa   
4 Americas 
5 Oceania

Here we get a tibble back with the unique continent names in their order of appearance. There is an equivalent base R way to do this with the unique() function:

# -------------------------------------------------------------------------
# Base R approach to determining the distinct continents

unique(gm97$continent)

[1] "Asia"     "Europe"   "Africa"   "Americas" "Oceania"

Either way, if there was a misspelled entry, for example, “Urope” we would have immediately seen it.

Checkpoint

Sorting data with `arrange()`

We have a rich dataset whose default ordering is by country. But it’s natural to ask questions like “What country has the highest life expectancy?” or “What country had the highest GDP per capita?” The arrange() function can help us quickly find the answers.

# -------------------------------------------------------------------------
# Sort the data by life expectancy
# NOTE: Default is ascending order

arrange(gm97, lifeExp)

# A tibble: 142 × 6
   country          year      pop continent lifeExp gdpPercap
   <chr>           <dbl>    <dbl> <chr>       <dbl>     <dbl>
 1 Rwanda           1997  7212583 Africa       36.1      590.
 2 Sierra Leone     1997  4578212 Africa       39.9      575.
 3 Zambia           1997  9417789 Africa       40.2     1071.
 4 Angola           1997  9875024 Africa       41.0     2277.
 5 Afghanistan      1997 22227415 Asia         41.8      635.
 6 Liberia          1997  2200725 Africa       42.2      609.
 7 Congo Dem. Rep.  1997 47798986 Africa       42.6      312.
 8 Somalia          1997  6633514 Africa       43.8      931.
 9 Uganda           1997 21210254 Africa       44.6      817.
10 Guinea-Bissau    1997  1193708 Africa       44.9      797.
# ℹ 132 more rows

This seems to have done an ascending order on lifeExp by default. We have some options, one of which is to wrap the column of interest in the desc() function to indicate we want the order to be descending.

# -------------------------------------------------------------------------
# Sort the data by descending life expectancy

arrange(gm97, desc(lifeExp))

# A tibble: 142 × 6
   country          year       pop continent lifeExp gdpPercap
   <chr>           <dbl>     <dbl> <chr>       <dbl>     <dbl>
 1 Japan            1997 125956499 Asia         80.7    28817.
 2 Hong Kong China  1997   6495918 Asia         80      28378.
 3 Sweden           1997   8897619 Europe       79.4    25267.
 4 Switzerland      1997   7193761 Europe       79.4    32135.
 5 Iceland          1997    271192 Europe       79.0    28061.
 6 Australia        1997  18565243 Oceania      78.8    26998.
 7 Italy            1997  57479469 Europe       78.8    24675.
 8 Spain            1997  39855442 Europe       78.8    20445.
 9 France           1997  58623428 Europe       78.6    25890.
10 Canada           1997  30305843 Americas     78.6    28955.
# ℹ 132 more rows

And here we see that Japan is the country with the longest lived citizens at a little over 80 years.

Checkpoint

Exercise

Which country has the highest GDP per capita in 1997?

arrange(gm97, desc(gdpPercap))

# A tibble: 142 × 6
   country        year       pop continent lifeExp gdpPercap
   <chr>         <dbl>     <dbl> <chr>       <dbl>     <dbl>
 1 Norway         1997   4405672 Europe       78.3    41283.
 2 Kuwait         1997   1765345 Asia         76.2    40301.
 3 United States  1997 272911760 Americas     76.8    35767.
 4 Singapore      1997   3802309 Asia         77.2    33519.
 5 Switzerland    1997   7193761 Europe       79.4    32135.
 6 Netherlands    1997  15604464 Europe       78.0    30246.
 7 Denmark        1997   5283663 Europe       76.1    29804.
 8 Austria        1997   8069876 Europe       77.5    29096.
 9 Canada         1997  30305843 Americas     78.6    28955.
10 Japan          1997 125956499 Asia         80.7    28817.
# ℹ 132 more rows

Subset columns with `select()`

Sometimes data tables have many columns, and it can be useful to select only a few of them to export, use downstream, preview, etc. The select() function works on columns:

# -------------------------------------------------------------------------
# Select country, year, and lifeExp columns from the table

select(gm97, country, year, lifeExp)

# A tibble: 142 × 3
   country      year lifeExp
   <chr>       <dbl>   <dbl>
 1 Afghanistan  1997    41.8
 2 Albania      1997    73.0
 3 Algeria      1997    69.2
 4 Angola       1997    41.0
 5 Argentina    1997    73.3
 6 Australia    1997    78.8
 7 Austria      1997    77.5
 8 Bahrain      1997    73.9
 9 Bangladesh   1997    59.4
10 Belgium      1997    77.5
# ℹ 132 more rows

We can do the equivalent of select() statement using base R with:

# -------------------------------------------------------------------------
# Base R appraoch to selecting columns from the table

gm97[ , c('country', 'year', 'lifeExp')]

# A tibble: 142 × 3
   country      year lifeExp
   <chr>       <dbl>   <dbl>
 1 Afghanistan  1997    41.8
 2 Albania      1997    73.0
 3 Algeria      1997    69.2
 4 Angola       1997    41.0
 5 Argentina    1997    73.3
 6 Australia    1997    78.8
 7 Austria      1997    77.5
 8 Bahrain      1997    73.9
 9 Bangladesh   1997    59.4
10 Belgium      1997    77.5
# ℹ 132 more rows

Here we use the square-bracket notation again, leaving the row position blank returns all rows, and in the column position we specify the names of the columns as a vector with the c() function, meaning “combine”, creates a temporary vector of column names. Note that we quoted the column names in this code, whereas in select() we didn’t have to.

Checkpoint

Columns can be removed using select() with a “minus” in front of the column name.

# -------------------------------------------------------------------------
# Remove the year column

select(gm97, -year)

# A tibble: 142 × 5
   country           pop continent lifeExp gdpPercap
   <chr>           <dbl> <chr>       <dbl>     <dbl>
 1 Afghanistan  22227415 Asia         41.8      635.
 2 Albania       3428038 Europe       73.0     3193.
 3 Algeria      29072015 Africa       69.2     4797.
 4 Angola        9875024 Africa       41.0     2277.
 5 Argentina    36203463 Americas     73.3    10967.
 6 Australia    18565243 Oceania      78.8    26998.
 7 Austria       8069876 Europe       77.5    29096.
 8 Bahrain        598561 Asia         73.9    20292.
 9 Bangladesh  123315288 Asia         59.4      973.
10 Belgium      10199787 Europe       77.5    27561.
# ℹ 132 more rows

Notice the year column is no longer displayed.

Saving objects in session

If we view gm97 we notice haven’t changed the original data because we haven’t saved the object.

# -------------------------------------------------------------------------
# Display gm97

gm97

# A tibble: 142 × 6
   country      year       pop continent lifeExp gdpPercap
   <chr>       <dbl>     <dbl> <chr>       <dbl>     <dbl>
 1 Afghanistan  1997  22227415 Asia         41.8      635.
 2 Albania      1997   3428038 Europe       73.0     3193.
 3 Algeria      1997  29072015 Africa       69.2     4797.
 4 Angola       1997   9875024 Africa       41.0     2277.
 5 Argentina    1997  36203463 Americas     73.3    10967.
 6 Australia    1997  18565243 Oceania      78.8    26998.
 7 Austria      1997   8069876 Europe       77.5    29096.
 8 Bahrain      1997    598561 Asia         73.9    20292.
 9 Bangladesh   1997 123315288 Asia         59.4      973.
10 Belgium      1997  10199787 Europe       77.5    27561.
# ℹ 132 more rows

To manipulate gm97 and save it as a new object we have to assign it a new object name:

# -------------------------------------------------------------------------
# Create a new table by selecting columns of an existing table

gm97_subset = select(gm97, country, year, lifeExp)
gm97_subset

# A tibble: 142 × 3
   country      year lifeExp
   <chr>       <dbl>   <dbl>
 1 Afghanistan  1997    41.8
 2 Albania      1997    73.0
 3 Algeria      1997    69.2
 4 Angola       1997    41.0
 5 Argentina    1997    73.3
 6 Australia    1997    78.8
 7 Austria      1997    77.5
 8 Bahrain      1997    73.9
 9 Bangladesh   1997    59.4
10 Belgium      1997    77.5
# ℹ 132 more rows

Saving objects to file

The original data file we read in data/gapminder_1997.csv remains the same throughout all these manipulations. What happens in R stays in R until we explicitly write a new object to the same file. It’s good practice to keep raw input data separate, and to never overwrite it. Let’s use the write_csv() function to write the gm97_subset data to a file. First let’s look at ?write_csv and determine what are the required parameters.

It looks like x, the data to be written and file, the path to the output file are required. There are other options, but we won’t use them here.

# -------------------------------------------------------------------------
# Write the subsetted table to file

write_csv(gm97_subset, file = 'gm97_subset.csv')

Checkpoint

Narrow down rows with `filter()`

We have seen that select() subsets the columns of a table. The function filter() subsets the rows of a data table based on logical criteria. So first, a note about logical operators in R:

Operator	Description	Example
<	less than	`pop < 575990`
<=	less than or equal to	`pop <= 575990`
>	greater than	`pop > 1000000`
>=	greater than or equal to	`pop >= 38000000`
==	exactly equal to	`continent == 'Africa'`
!=	not equal to	`continent != 'Asia'`
!x	not x	`!(continent == 'Africa')`
a \| b	a or b	`pop < 575990 \| pop > 1000000`
a & b	a and b	`continent == 'Asia' & continent == 'Africa'`
a %in% b	a in b	`continent %in% c('Asia', 'Africa')`

With these operators in mind we can filter the gm97 data table in a variety of ways. To filter for only the African data we could:

# -------------------------------------------------------------------------
# Filter the data for only countries in Africa

filter(gm97, continent == 'Africa')

# A tibble: 52 × 6
   country                   year      pop continent lifeExp gdpPercap
   <chr>                    <dbl>    <dbl> <chr>       <dbl>     <dbl>
 1 Algeria                   1997 29072015 Africa       69.2     4797.
 2 Angola                    1997  9875024 Africa       41.0     2277.
 3 Benin                     1997  6066080 Africa       54.8     1233.
 4 Botswana                  1997  1536536 Africa       52.6     8647.
 5 Burkina Faso              1997 10352843 Africa       50.3      946.
 6 Burundi                   1997  6121610 Africa       45.3      463.
 7 Cameroon                  1997 14195809 Africa       52.2     1694.
 8 Central African Republic  1997  3696513 Africa       46.1      741.
 9 Chad                      1997  7562011 Africa       51.6     1005.
10 Comoros                   1997   527982 Africa       60.7     1174.
# ℹ 42 more rows

The base R equivalent of the above uses the square bracket notation that we saw earlier with a twist:

# -------------------------------------------------------------------------
# Base R approach to filtering rows

gm97[gm97$continent == 'Africa', ]

# A tibble: 52 × 6
   country                   year      pop continent lifeExp gdpPercap
   <chr>                    <dbl>    <dbl> <chr>       <dbl>     <dbl>
 1 Algeria                   1997 29072015 Africa       69.2     4797.
 2 Angola                    1997  9875024 Africa       41.0     2277.
 3 Benin                     1997  6066080 Africa       54.8     1233.
 4 Botswana                  1997  1536536 Africa       52.6     8647.
 5 Burkina Faso              1997 10352843 Africa       50.3      946.
 6 Burundi                   1997  6121610 Africa       45.3      463.
 7 Cameroon                  1997 14195809 Africa       52.2     1694.
 8 Central African Republic  1997  3696513 Africa       46.1      741.
 9 Chad                      1997  7562011 Africa       51.6     1005.
10 Comoros                   1997   527982 Africa       60.7     1174.
# ℹ 42 more rows

Here we put the condition in the row position, before the comma, because we want to subset rows. The blank after the comma indicates we want all the columns returned.

Exercise

How can we filter the data for just the United Kingdom?

filter(gm97, country == 'United Kingdom')

# A tibble: 1 × 6
  country         year      pop continent lifeExp gdpPercap
  <chr>          <dbl>    <dbl> <chr>       <dbl>     <dbl>
1 United Kingdom  1997 58808266 Europe       77.2    26075.

We can answer the question, “Which African countries have population over 10,000,000?” by using the & operator with our previous code. This requires both conditions to be true at the same time:

# -------------------------------------------------------------------------
# Filter the data for African countries with populations more than 10M.

filter(gm97, continent == 'Africa' & pop >= 10000000)

# A tibble: 19 × 6
   country          year       pop continent lifeExp gdpPercap
   <chr>           <dbl>     <dbl> <chr>       <dbl>     <dbl>
 1 Algeria          1997  29072015 Africa       69.2     4797.
 2 Burkina Faso     1997  10352843 Africa       50.3      946.
 3 Cameroon         1997  14195809 Africa       52.2     1694.
 4 Congo Dem. Rep.  1997  47798986 Africa       42.6      312.
 5 Cote d'Ivoire    1997  14625967 Africa       48.0     1786.
 6 Egypt            1997  66134291 Africa       67.2     4173.
 7 Ethiopia         1997  59861301 Africa       49.4      516.
 8 Ghana            1997  18418288 Africa       58.6     1005.
 9 Kenya            1997  28263827 Africa       54.4     1360.
10 Madagascar       1997  14165114 Africa       55.0      986.
11 Malawi           1997  10419991 Africa       47.5      692.
12 Morocco          1997  28529501 Africa       67.7     2982.
13 Mozambique       1997  16603334 Africa       46.3      472.
14 Nigeria          1997 106207839 Africa       47.5     1625.
15 South Africa     1997  42835005 Africa       60.2     7479.
16 Sudan            1997  32160729 Africa       55.4     1632.
17 Tanzania         1997  30686889 Africa       48.5      789.
18 Uganda           1997  21210254 Africa       44.6      817.
19 Zimbabwe         1997  11404948 Africa       46.8      792.

We can subset the data for very large or very small population with the | operator, which requires either condition to be true:

# -------------------------------------------------------------------------
# Filter the data for high and low populations

filter(gm97, pop <= 1000000 | pop >= 1000000000)

# A tibble: 9 × 6
  country                year        pop continent lifeExp gdpPercap
  <chr>                 <dbl>      <dbl> <chr>       <dbl>     <dbl>
1 Bahrain                1997     598561 Asia         73.9    20292.
2 China                  1997 1230075000 Asia         70.4     2289.
3 Comoros                1997     527982 Africa       60.7     1174.
4 Djibouti               1997     417908 Africa       53.2     1895.
5 Equatorial Guinea      1997     439971 Africa       48.2     2814.
6 Iceland                1997     271192 Europe       79.0    28061.
7 Montenegro             1997     692651 Europe       75.4     6466.
8 Reunion                1997     684810 Africa       74.8     6072.
9 Sao Tome and Principe  1997     145608 Africa       63.3     1339.

This gives us all the countries with fewer than a million people or more than a billion people.

Checkpoint

Make new columns with `mutate()`

It’s common to use existing columns of a data table to create new columns. For example, in the gm97 data we have a pop column and a gdpPercap column. We could multiply these two columns to get a new gdp column. The dplyr function mutate() adds columns to an existing data table.

# -------------------------------------------------------------------------
# Compute the GDP of each country by multiplying GDP per capita and population

mutate(gm97, gdp = pop * gdpPercap)

# A tibble: 142 × 7
   country      year       pop continent lifeExp gdpPercap           gdp
   <chr>       <dbl>     <dbl> <chr>       <dbl>     <dbl>         <dbl>
 1 Afghanistan  1997  22227415 Asia         41.8      635.  14121995875.
 2 Albania      1997   3428038 Europe       73.0     3193.  10945912519.
 3 Algeria      1997  29072015 Africa       69.2     4797. 139467033682.
 4 Angola       1997   9875024 Africa       41.0     2277.  22486820881.
 5 Argentina    1997  36203463 Americas     73.3    10967. 397053586287.
 6 Australia    1997  18565243 Oceania      78.8    26998. 501223252921.
 7 Austria      1997   8069876 Europe       77.5    29096. 234800471832.
 8 Bahrain      1997    598561 Asia         73.9    20292.  12146009862.
 9 Bangladesh   1997 123315288 Asia         59.4      973. 119957417048.
10 Belgium      1997  10199787 Europe       77.5    27561. 281118335091.
# ℹ 132 more rows

As before, this didn’t actually add the column to gm97, we have to save the object to store it:

# -------------------------------------------------------------------------
# Now save the updated table as a new object

gm97_gdp = mutate(gm97, gdp = pop * gdpPercap)
gm97_gdp

# A tibble: 142 × 7
   country      year       pop continent lifeExp gdpPercap           gdp
   <chr>       <dbl>     <dbl> <chr>       <dbl>     <dbl>         <dbl>
 1 Afghanistan  1997  22227415 Asia         41.8      635.  14121995875.
 2 Albania      1997   3428038 Europe       73.0     3193.  10945912519.
 3 Algeria      1997  29072015 Africa       69.2     4797. 139467033682.
 4 Angola       1997   9875024 Africa       41.0     2277.  22486820881.
 5 Argentina    1997  36203463 Americas     73.3    10967. 397053586287.
 6 Australia    1997  18565243 Oceania      78.8    26998. 501223252921.
 7 Austria      1997   8069876 Europe       77.5    29096. 234800471832.
 8 Bahrain      1997    598561 Asia         73.9    20292.  12146009862.
 9 Bangladesh   1997 123315288 Asia         59.4      973. 119957417048.
10 Belgium      1997  10199787 Europe       77.5    27561. 281118335091.
# ℹ 132 more rows

Checkpoint

Combine functions with the `%>%` pipe

We’ve learned how to subset rows with filter(), select columns with select(), get distinct elements with distinct(), and summarize columns with summarize(). It’s common to use these functions in sequence, where the output of one becomes the input of the next. The concept of “pipe” (|) from bash has an equivalent in the tidyverse in the %>% symbol. To see this in action, let’s select data from Oceania and display only the country, continent, and pop columns.

# -------------------------------------------------------------------------
# Filter for countries in Oceania, then show a subset of the columns

gm97 %>% filter(continent == 'Oceania') %>% select(country, continent, pop)

# A tibble: 2 × 3
  country     continent      pop
  <chr>       <chr>        <dbl>
1 Australia   Oceania   18565243
2 New Zealand Oceania    3676187

Question

Does the order of filter() and select() matter? Would something like the following work, why or why not?
gm97 %>% 
    select(country, pop) %>% 
    filter(continent == 'Oceania')

It seems there are only two countries, Australia and New Zealand. If we wanted to take the mean of the populations we could pipe the data, after filtering, to summarize() as in:

# -------------------------------------------------------------------------
# Filter for countries in Oceania and then compute the mean population

gm97 %>%
    filter(continent == 'Oceania') %>%
    summarize(mean_pop = mean(pop))

# A tibble: 1 × 1
  mean_pop
     <dbl>
1 11120715

Note this is distinct from taking the mean over the entire dataset:

# -------------------------------------------------------------------------
# Use the same summarize() statement, but on the entire dataset

gm97 %>% summarize(mean_pop = mean(pop))

# A tibble: 1 × 1
   mean_pop
      <dbl>
1 38839468.

Here we’ve meaningfully named our output, but that was optional. For quick data exploration that could be left off.

Checkpoint

Group rows with `group_by()`

If we wanted to determine the mean populations over all the continents, we could run the above code we used for Oceania, but replace Oceania with each continent name. That’s a bit tedious, and the tidyverse provides a very nice function, group_by(), that breaks a data table up into groups and will perform operations within the groups.

# -------------------------------------------------------------------------
# Compute the mean population per continent

gm97 %>% 
    group_by(continent) %>% 
    summarize(mean_pop = mean(pop))

# A tibble: 5 × 2
  continent   mean_pop
  <chr>          <dbl>
1 Africa     14304480.
2 Americas   31876016.
3 Asia      102523803.
4 Europe     18964805.
5 Oceania    11120715

Notice that the Oceania average matches what we got above. Going back to our Oceania subset, we can use group_by() with the counting function n() to count how many countries there are per continent in the gm97 data, and verify that there really only are two:

# -------------------------------------------------------------------------
# Determine the number of countries on each continent in the dataset

gm97 %>% 
    group_by(continent) %>% 
    summarize(num_countries_per_continent = n())

# A tibble: 5 × 2
  continent num_countries_per_continent
  <chr>                           <int>
1 Africa                             52
2 Americas                           25
3 Asia                               33
4 Europe                             30
5 Oceania                             2

There is a base R function, table(), that tabulates the number of time an entry occurs in a vector, and is useful for quick sanity checking:

# -------------------------------------------------------------------------
# Base R approach for determining the number of countries per continent

table(gm97$continent)


  Africa Americas     Asia   Europe  Oceania 
      52       25       33       30        2

Objectives

Learn essential data manipulation functions of dplyr:
- select() columns
- filter() rows
- arrange() rows
- mutate() columns
- group_by() rows
Learn about the pipe operator (%>%) and how to use it to chain together multiple functions.

Exercise

In summarizing data, it’s often useful to build a nicely formatted table with min, median, and max values of a variable. Below is an example of code to do that for the populations by continent, and arranged by the median population.
gm97 %>% 
    group_by(continent) %>%
    summarize(
        min_POP = min(pop),
        median_POP = median(pop),
        max_POP = max(pop)
    ) %>% 
    arrange(median_POP)
# A tibble: 5 × 4
  continent min_POP median_POP    max_POP
  <chr>       <dbl>      <dbl>      <dbl>
1 Africa     145608   7805422.  106207839
2 Americas  1138101   7992357   272911760
3 Europe     271192   9527017    82011073
4 Oceania   3676187  11120715    18565243
5 Asia       598561  21229759  1230075000
Try to write the code that would generate the following table, similar to the above, but for the GDP per continent instead of the population, including the mean GDP in addition to the min, median, and max, as well as ordering the results by mean GDP.
# A tibble: 5 × 5
  continent      min_GDP      mean_GDP    median_GDP max_GDP
  <chr>            <dbl>         <dbl>         <dbl>   <dbl>
1 Africa      194980183.  30023173824.   8190364646. 3.20e11
2 Oceania   77385257446. 289304255183. 289304255183. 5.01e11
3 Europe     4478413552. 383606933833. 172462904126. 2.28e12
4 Asia       4745744246. 387597655323. 105396057842. 3.63e12
5 Americas   9276089515. 582693307146.  45924427025. 9.76e12

Previous lesson	Top of this lesson	Next lesson

Data Transformation with dplyr

Objectives

Loading data

Exploring data tables

Get stats with summary()

Get stats with summarize()

Tip: Reference columns by name

Find unique values with distinct()

Sorting data with arrange()

Exercise

Subset columns with select()

Saving objects in session

Saving objects to file

Narrow down rows with filter()

Exercise

Make new columns with mutate()

Combine functions with the %>% pipe

Question

Group rows with group_by()

Objectives

Exercise

Get stats with `summary()`

Get stats with `summarize()`

Find unique values with `distinct()`

Sorting data with `arrange()`

Subset columns with `select()`

Narrow down rows with `filter()`

Make new columns with `mutate()`

Combine functions with the `%>%` pipe

Group rows with `group_by()`