Exploring the count() Function in R’s tidyverse

R
tidyverse
regex
Tips and Tricks in Data Cleaning and Visualization, Part III.
Author

Carlos Fernández

Published

June 3, 2024

R’s tidyverse is a collection of packages designed for data science. One of the most useful functions within this suite is count(), part of the dplyr package. This function is used to count occurrences of unique values within a dataset, which is crucial for data analysis and exploratory data analysis (EDA). In this blog post, we’ll delve into how count() works, particularly with the sort = TRUE option, and we’ll use datasets from the medicaldata package to illustrate its application.

Introduction to count()

The count() function in dplyr is designed to count the number of occurrences of each unique value of one or more variables. It returns a data frame with the counts of these values. Here’s a basic syntax of the function:

count(data, vars, wt = NULL, sort = FALSE, name = "n")

  • data: The data frame.
  • vars: The variable(s) to count unique values of.
  • wt: Optional. If provided, counts will be weighted by this variable.
  • sort: If TRUE, the resulting data frame will be sorted in descending order by the count.
  • name: The name of the count column. Default is “n”.

Using count() with sort = TRUE

Setting sort = TRUE orders the output by the count in descending order, which is helpful when you want to quickly see the most frequent values.

Example with the medicaldata package

The medicaldata package contains various datasets from medical research, which are great for demonstration. We’ll use the covid_testing dataset for our example.

First, install and load the required packages:

Show the code 
# install.packages("tidyverse")
# install.packages("medicaldata")

library(tidyverse)
library(medicaldata)

Next, load the covid_testing dataset and take a look at its structure:

Show the code 
data("covid_testing")
glimpse(covid_testing)
Rows: 15,524
Columns: 17
$ subject_id      <dbl> 1412, 533, 9134, 8518, 8967, 11048, 663, 2158, 3794, 4…
$ fake_first_name <chr> "jhezane", "penny", "grunt", "melisandre", "rolley", "…
$ fake_last_name  <chr> "westerling", "targaryen", "rivers", "swyft", "karstar…
$ gender          <chr> "female", "female", "male", "female", "male", "female"…
$ pan_day         <dbl> 4, 7, 7, 8, 8, 8, 9, 9, 9, 9, 9, 9, 9, 9, 9, 10, 10, 1…
$ test_id         <chr> "covid", "covid", "covid", "covid", "covid", "covid", …
$ clinic_name     <chr> "inpatient ward a", "clinical lab", "clinical lab", "c…
$ result          <chr> "negative", "negative", "negative", "negative", "negat…
$ demo_group      <chr> "patient", "patient", "patient", "patient", "patient",…
$ age             <dbl> 0.0, 0.0, 0.8, 0.8, 0.8, 0.8, 0.8, 0.0, 0.0, 0.9, 0.9,…
$ drive_thru_ind  <dbl> 0, 1, 1, 1, 0, 0, 1, 0, 1, 1, 1, 0, 0, 1, 1, 1, 1, 0, …
$ ct_result       <dbl> 45, 45, 45, 45, 45, 45, 45, 45, 45, 45, 45, 45, 45, 45…
$ orderset        <dbl> 0, 0, 1, 1, 1, 0, 1, 1, 1, 1, 0, 1, 0, 1, 1, 1, 1, 1, …
$ payor_group     <chr> "government", "commercial", NA, NA, "government", "com…
$ patient_class   <chr> "inpatient", "not applicable", NA, NA, "emergency", "r…
$ col_rec_tat     <dbl> 1.4, 2.3, 7.3, 5.8, 1.2, 1.4, 2.6, 0.7, 1.0, 7.1, 2.5,…
$ rec_ver_tat     <dbl> 5.2, 5.8, 4.7, 5.0, 6.4, 7.0, 4.2, 6.3, 5.6, 7.0, 3.8,…

The covid_testing dataset contains data from deidentified results of COVID-19 testing at the Children’s Hospital of Pennsylvania (CHOP) in 2020. Suppose we want to count the number of participants according to the COVID-19 test result and sort the results by the count in descending order. Here’s how we can do that:

Show the code 
covid_testing %>%
  count(result, sort = TRUE)
# A tibble: 3 × 2
  result       n
  <chr>    <int>
1 negative 14358
2 positive   865
3 invalid    301

Interpreting the Results

The output is a tibble where the first column is the test result, and the second column, named n, shows the counts. Because we used sort = TRUE, the result with the highest count (negative) will appear first (n = 14,358 patients), followed by positive results (n = 865), and invalid results (n = 301).

Adding Multiple Variables

You can also count combinations of multiple variables. For example, if we want to count combinations of result and gender:

Show the code 
covid_testing %>%
  count(result, gender, sort = TRUE)
# A tibble: 6 × 3
  result   gender     n
  <chr>    <chr>  <int>
1 negative female  7237
2 negative male    7121
3 positive female   449
4 positive male     416
5 invalid  male     155
6 invalid  female   146

This will give us a data frame with the counts of each result-gender combination, sorted by the count.

Practical Use Cases

The count() function is particularly useful for:

  • EDA: Quickly understanding the distribution of values in your dataset.
  • Data Cleaning: Identifying and handling rare or common categories.
  • Reporting: Summarizing data in a clear and sorted manner for presentations or reports.

Conclusion

The count() function in R’s tidyverse is a powerful tool for summarizing data. By using the sort = TRUE option, you can quickly identify the most frequent values in your dataset. Whether you’re counting single variables or combinations of variables, count() simplifies the task and makes your data analysis workflow more efficient.

Happy counting!

References

  • Higgins P (2021). medicaldata: Data Package for Medical Datasets. R package version 0.2.0, https://CRAN.R-project.org/package=medicaldata.

  • Wickham H, Averick M, Bryan J, Chang W, McGowan LD, François R, Grolemund G, Hayes A, Henry L, Hester J, Kuhn M, Pedersen TL, Miller E, Bache SM, Müller K, Ooms J, Robinson D, Seidel DP, Spinu V, Takahashi K, Vaughan D, Wilke C, Woo K, Yutani H (2019). “Welcome to the tidyverse.” Journal of Open Source Software, 4(43), 1686. doi:10.21105/joss.01686 https://doi.org/10.21105/joss.01686.