R Challenge #1: World Population

I'm proud to say I've joined the esteemed ranks of Code Challenge Authors. In this and a set of future articles, I'll share some backstory behind each challenge. My peers have supplied other code challenges you might find interesting: Javascript ... Python ... Java ... Github ... HTML ... SQL ... SQL for Data Science ... PHP

R Challenge #1 : Import the World Population database

Importing CSV into an R data object would seem straightforward - but you quickly run into nuances that will foul your data. Data types, factors, missing data, incomplete lines and more cause importing to be a nightmare. In this episode, I challenge you to import a CSV file.

Here's My Solution

# import the United Nations world population database
# all fields should be integer or numeric
# except variant = factor, location = character


worldPop <- read.csv("https://population.un.org/wpp/Download/Files/1_Indicators%20(Standard)/CSV_FILES/WPP2019_TotalPopulationBySex.csv",
? ? ? ? ?colClasses = c("integer","character","integer", "factor", "integer", "numeric","numeric","numeric","numeric","numeric"))

        

What You Don't See

Writing courses for LinkedIn (and other online learning portals) requires sample code to be both public domain and accessible - AND - it should be interesting to work with. The UN world population data fills all three, but MAN, it was difficult to find. I spent more time looking for the dataset than I did actually writing the code.

An option would be to download the data and include it with the example files. Which would contribute to a multi-gigabyte download. Some of you don't have that kind of patience and bandwidth, so it's easier to grab it directly from the code.

An Alternate Solution

I have an ongoing debate with my peers regarding base R vs tidy R. Should I teach the tidy verse (read_csv)? Or should I teach base R (read.csv)? So far, I've focused on base R - but I'm aware there are cleaner solutions available in the tidyverse. For example, compare this code to the above.

library(readr)

WPP2019_TotalPopulationBySex <- read_csv("https://population.un.org/wpp/Download/Files/1_Indicators%20(Standard)/CSV_FILES/WPP2019_TotalPopulationBySex.csv",
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?col_types = "icifnnnnnn")
        

It looks similar, but read_csv is faster and makes better decisions about factors and missing data. Plus it creates an object tailored to other routines in the tidyverse.

Choosing a Problem

Choosing data is one aspect of authoring a course. Choosing a problem is another. Good challenges require some thought, possibly resulting in failure the first time through. I believe it's possible to import the data within two or three attempts.

And...good challenges are focused. I hope to illustrate one concept - maybe two. If a challenge requires multiple solutions, an early failure obscures the remaining lessons. Less complex challenges are easier to focus on a concept - but they may be less interesting to solve.

How about you?

Do you have some opinions on this challenge? Please share them in the comments below.

mnr