widyr: Widen, process, and re-tidy a dataset

David Robinson

2017-08-13

This package wraps the pattern of un-tidying data into a wide matrix, performing some processing, then turning it back into a tidy form. This is useful for several mathematical operations such as co-occurence counts, correlations, or clustering that are best done on a wide matrix.

Towards a precise definition of “wide” data

The term “wide data” has gone out of fashion as being “imprecise” (Wickham 2014)), but I think with a proper definition the term could be entirely meaningful and useful.

A wide dataset is one or more matrices where:

When would you want data to be wide rather than tidy? Notable examples include classification, clustering, correlation, factorization, or other operations that can take advantage of a matrix structure. In general, when you want to compare between items rather than compare between variables, this is a useful structure.

The widyr package is based on the observation that during a tidy data analysis, you often want data to be wide only temporarily, before returning to a tidy structure for visualization and further analysis. widyr makes this easy through a set of pairwise_ functions.

Example: gapminder

Consider the gapminder dataset in the gapminder package.

library(dplyr)
library(gapminder)

gapminder
## # A tibble: 1,704 x 6
##        country continent  year lifeExp      pop gdpPercap
##         <fctr>    <fctr> <int>   <dbl>    <int>     <dbl>
##  1 Afghanistan      Asia  1952  28.801  8425333  779.4453
##  2 Afghanistan      Asia  1957  30.332  9240934  820.8530
##  3 Afghanistan      Asia  1962  31.997 10267083  853.1007
##  4 Afghanistan      Asia  1967  34.020 11537966  836.1971
##  5 Afghanistan      Asia  1972  36.088 13079460  739.9811
##  6 Afghanistan      Asia  1977  38.438 14880372  786.1134
##  7 Afghanistan      Asia  1982  39.854 12881816  978.0114
##  8 Afghanistan      Asia  1987  40.822 13867957  852.3959
##  9 Afghanistan      Asia  1992  41.674 16317921  649.3414
## 10 Afghanistan      Asia  1997  41.763 22227415  635.3414
## # ... with 1,694 more rows

This tidy format (one-row-per-country-per-year) is very useful for grouping, summarizing, and filtering operations. But if we want to compare pairs of countries (for example, to find countries that are similar to each other), we would have to reshape this dataset. Note that here, country is the item, while year is the feature column.

Pairwise operations

The widyr package offers pairwise_ functions that operate on pairs of items within data. An example is pairwise_dist:

library(widyr)

gapminder %>%
  pairwise_dist(country, year, lifeExp)
## # A tibble: 20,022 x 3
##         item1       item2  distance
##        <fctr>      <fctr>     <dbl>
##  1    Albania Afghanistan 107.41825
##  2    Algeria Afghanistan  76.75286
##  3     Angola Afghanistan   4.64934
##  4  Argentina Afghanistan 109.50686
##  5  Australia Afghanistan 128.95745
##  6    Austria Afghanistan 123.51771
##  7    Bahrain Afghanistan  98.13426
##  8 Bangladesh Afghanistan  45.33990
##  9    Belgium Afghanistan 125.41156
## 10      Benin Afghanistan  39.32262
## # ... with 20,012 more rows

In a single step, this finds the Euclidean distance between the lifeExp value in each pair of countries, matching pairs based on year. We could find the closest pairs of countries overall with arrange():

gapminder %>%
  pairwise_dist(country, year, lifeExp) %>%
  arrange(distance)
## # A tibble: 20,022 x 3
##             item1          item2 distance
##            <fctr>         <fctr>    <dbl>
##  1        Germany        Belgium 1.075702
##  2        Belgium        Germany 1.075702
##  3 United Kingdom    New Zealand 1.509025
##  4    New Zealand United Kingdom 1.509025
##  5         Norway    Netherlands 1.557933
##  6    Netherlands         Norway 1.557933
##  7          Italy         Israel 1.662690
##  8         Israel          Italy 1.662690
##  9        Finland        Austria 1.936558
## 10        Austria        Finland 1.936558
## # ... with 20,012 more rows

Notice that this includes duplicates (Germany/Belgium and Belgium/Germany). To avoid those (the upper triangle of the distance matrix), use upper = FALSE:

gapminder %>%
  pairwise_dist(country, year, lifeExp, upper = FALSE) %>%
  arrange(distance)
## # A tibble: 10,011 x 3
##          item1          item2 distance
##         <fctr>         <fctr>    <dbl>
##  1     Belgium        Germany 1.075702
##  2 New Zealand United Kingdom 1.509025
##  3 Netherlands         Norway 1.557933
##  4      Israel          Italy 1.662690
##  5     Austria        Finland 1.936558
##  6     Belgium United Kingdom 1.949243
##  7     Iceland         Sweden 2.005176
##  8     Comoros     Mauritania 2.008199
##  9     Belgium  United States 2.092081
## 10     Germany        Ireland 2.097239
## # ... with 10,001 more rows

In some analyses, we may be interested in correlation rather than distance of pairs. For this we would use pairwise_cor:

gapminder %>%
  pairwise_cor(country, year, lifeExp, upper = FALSE, sort = TRUE)
## # A tibble: 10,011 x 3
##           item1                 item2 correlation
##          <fctr>                <fctr>       <dbl>
##  1    Indonesia            Mauritania   0.9996291
##  2      Morocco               Senegal   0.9995515
##  3 Saudi Arabia    West Bank and Gaza   0.9995156
##  4       Brazil                France   0.9994246
##  5      Bahrain               Reunion   0.9993649
##  6     Malaysia Sao Tome and Principe   0.9993233
##  7         Peru                 Syria   0.9993063
##  8      Bolivia                Gambia   0.9992930
##  9    Indonesia               Morocco   0.9992799
## 10        Libya               Senegal   0.9992710
## # ... with 10,001 more rows