Extra practice exercises

Overview

Teaching: 0 min
Exercises: 0 min

Questions

• How to apply the tools and concepts learned to new data?

Objectives

• Use R and the tidyverse collection of packages to explore a new dataset.

This document suggests new exercises to practice the skills and concepts learned in the course. Each exercise section links to one or several of the main lesson episodes it refers to.

(Note: solutions are not yet available for these exercises and will be added in time)

Starting a data analysis project

This exercise uses skills learned in the episodes Introduction to R and RStudio, Basic objects and data types in R and Working with Tabular Data.

For this set of questions, we will use a new dataset from the Gapminder Foundation, which includes data about energy consumption and CO2 emmissions in the different countries. This new data file includes the following variables:

Part I - setting up your project

1. On your computer, create a new directory for this project, for example gapminder-energy
2. Create a new R Project on this directory
3. Create directories that you feel are important, including one for saving the raw data.
4. Download the gapminder energy dataset and save it in your project’s data folder. (right-click the file link and choose “Save link as…”)
5. Install the skimr package, which you will use to do some quality control checks on the data file. (Hint: install.packages())

Part II - reading and checking quality of data

1. Read the gapminder1990to2010_energy.tsv file into a data.frame/tibble object called energy. (Hint: note that this is a tab-delimited file.)
2. Make a basic check that your data import went well. Your data frame should have 3968 rows and 7 columns. (Hint: use summary() and str() to check data integrity)
3. Use the skim() function from the skimr package to get an informative summary of your data frame. Make sure columns were imported as the correct type of value, and that you have the expected number of categories for character columns. (Hint: don’t forget to load the package first with library())

Note any problems you find during this exploratory analysis.

Data Visualisation

Following from the previous exercise on the gapminder1990to2010_energy.tsv dataset:

1. Use a histogram to investigate the distribution of yearly CO2 emissions in the dataset. Consider whether the original scale is adequate.
2. Has the distribution of yearly CO2 emissions changed dramatically over the years? Try using geom_density() or geom_freqpoly() with the following aesthetic group = year. (to ensure one line is drawn for each year).
3. Look at the trend over time (x-axis) of CO2 emissions (y-axis) for each country, both on a linear and on a log scales. How does the scale change your perception of the data? (hint use geom_line() with aes(group = country)).
4. Try and recreate the following graph, showing the relationship between energy use and CO2 emissions. Note that the axis are on a log-scale. The “Dark2” palette was used from the brewer colour scale.

Data manipulation

The gapminder1990to2010_energy.tsv dataset contains some mistakes that need correcting.

Create a new table, energy_clean, that has the following issues fixed:

• Convert the energy_production_per_person from character to numeric. (hint: use the as.numeric() function within mutate())
• country has 191 unique categories, whereas country_id has 190 (you can check this with unique(energy$country) and unique(energy$country_id)). One of the countries, Brazil is spelled with an “s” and a “z”. Change the values of “Brasil” to “Brazil”. (hint: use the ifelse() function within mutate())

From the new table, try and fix the following code, by writing the correct condition in filter(), which produces the plot below:

energy_clean %>% 
  filter(FIXME) %>% 
  ggplot(aes(energy_use_per_person, energy_production_per_person)) +
  geom_point(aes(colour = world_region)) + 
  facet_wrap(facets = vars(year)) +
  scale_x_continuous(trans = "log10") +
  scale_y_continuous(trans = "log10") +
  scale_colour_brewer(palette = "Dark2")

Key Points

• The initial exploration of data is crucial to detect any data quality issues that need fixing.