For our homework exercises, we will use a new dataset from the Gapminder Foundation, which gives access to global data as well as many tools to help explore it.

We will use data relating to socio-economic statistics for 2010. The columns in our data file are:

Column Description
country country name
world_region 6 world regions
year year that each datapoint refers to
children_per_woman total fertility rate
life_expectancy average number of years a newborn child would live if current mortality patterns were to stay the same
income_per_person gross domestic product per person adjusted for differences in purchasing power
is_oecd Whether a country belongs to the “OECD” (TRUE) or not (FALSE)
income_groups categorical classification of income groups
population total number of a country’s population
main_religion religion of the majority of population in 2008
child_mortality death of children under 5 years old per 1000 births
life_expectancy_female life expectancy at birth, females
life_expectancy_male life expectancy at birth, males

Task 1 - setting up your project

Tip: this exercise builds on the skills gained in Getting Started.

  1. On your computer, create a new directory for this project called gapminder-dataviz.
  2. Create directories that you feel are important, including one called data_raw for saving the raw data.
  3. Download the gapminder socio-economic dataset 2010 and save it in your project’s data folder. (right-click the file link and choose “Save link as…”)
  4. Create a new R Project on the gapminder-dataviz directory you just created.
Answer

We create a new directory as well as sub-directories, shown here schematically:

gapminder-dataviz
 |_ data_raw
 |_ data_processed
 |_ fig_output
 |_ scripts

We use data_raw to save the data file that we download with the link provided.

Finally, we create an R project on this directory:

  1. Start RStudio.
  2. Under the File menu, click on New Project. Choose Existing Directory.
  3. Hit the Browse… button to go to the folder named gapminder-dataviz that we’ve just created.
  4. Click on Create Project.

RStudio should refresh itself and then indicate that the working directory has been set to the new folder. For example, you can run the command getwd() on the console to confirm that this is the case.

Task 2 - importing data

As with any dataset you must first understand its content and formatting. Understanding what data you have will help you decide what story you can learn from the data and how best to present it

Create a new script to analyse these data and call it 01-gapminder_exploration.R. Then, populate it with code to achieve the following:

  1. Read the gapminder2010_socioeconomic.csv file into a data.frame/tibble object called gapminder.
    Hint Use the read_csv() function. Remember to first load the tidyverse package with library(tidyverse).
  2. Make a basic check that your data import went well. Your data frame should have 193 rows and 13 columns.
    Hint Use functions such as nrow(), ncol(), summary() and str() to check data integrity.
  3. What types of variables does this data have?
  4. Are there any issues that you notice with these data? Note these down.
Answer

We can read our data as follows:

library(tidyverse)

gapminder <- read_csv("data_raw/gapminder2010_socioeconomic.csv")
#> Rows: 193 Columns: 13
#> ── Column specification ──────────────────
#> Delimiter: ","
#> chr (5): country, world_region, income_groups, main_religion, life_expectanc...
#> dbl (7): year, children_per_woman, life_expectancy, income_per_person, popul...
#> lgl (1): is_oecd
#> 
#> 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.

To examine the contents of the data.frame we can use several functions, for example, to get the number of rows and columns:

ncol(gapminder) # number of columns in the data.frame
nrow(gapminder) # number of rows in the data.frame

The str() function gives a more comprehensive view of the contents of the data.frame, including the number of rows and columns as well as the type of variable each columns was imported as by R:

str(gapminder)

The summary function is also very useful as it gives a quick overview of the types of variables as well as average and quantiles for numeric data:

summary(gapminder)

The types of variables we have are:

  • nominal (categorical variable with no order): for example country and world_region
  • ordinal (categorical variable with order): for example income_groups
  • binary (categorical variable with two mutually exclusive values): is_oecd
  • continuous (numerical variable with any values): for example income_per_person or life_expectancy
  • discrete (numerical variable with limited values only): for example year

From the output of the summary function, we can notice a few issues with these data:

  • Some life_expectancy_male values are invalid == -999.
  • life_expectancy_female was imported as a character variable, but should be numeric.
  • Some missing data - this is almost always expected in large datasets.

If we look at the top few rows of the table:

head(gapminder)

We can see that life_expectancy_female is showing the value “-” in the 4th row of data. Probably the person recording these data encoded missing values with the “-” symbol, but the read_csv() function did not recognise this as missing data. The default is to consider empty cells as missing data and so we should correct this in the dataset to make sure all missing values are encoded in the same way.

There are a few other issues in the main_religion column, which were a little harder to detect. If we look at the unique values of this column, we will notice different spellings/formats for some of its values:

# check unique values in a column
unique(gapminder$main_religion)

These types of spelling mistakes are very common and it’s important to be aware that R would consider “muslim” and “Muslim” to be different words (due to the case-sensitivity).

Task 3 - plotting

  1. Create a scatter plot for life expectancy (x) against income per person (y)
  2. You will notice you get a warning message here - why do you think this is?
Answer 
ggplot(data = gapminder, 
       mapping = aes(x = life_expectancy, y = income_per_person)) +
  geom_point()
#> Warning: Removed 6 rows containing missing values
#> (`geom_point()`).

The warning message we get is because 6 of the rows in the data frame do not have life_expectancy information (they are NA missing data).

If it was a key variable for your analysis you might have wanted to remove those individuals with missing data. In this case, we don’t mind having these missing data, so we can carry on with our analysis.

Note that ‘Warning messages’ are simply that - a warning, not an error. They are very helpful and always worth reading.

Task 4 - data manipulation

  1. How many countries are there in “South Asia”?
    Hint Use the filter() function to subset the table to retain only rows where world_region == "south_asia".
  2. Create a new column in the table called income_total, which is the product of population and income per person (i.e. the total average income of the country).
    Hint Use the mutate() function to create a new column.
Answer

To identify countries in South Asia, we can use the following:

gapminder %>% 
  filter(world_region == "south_asia")

From the output we can see this table has 8 rows, therefore 8 countries in this part of the world.

To create the new column we can use the mutate() function, as such:

gapminder %>% 
  mutate(income_total = population*income_per_person)

If we wanted to save this in our table, we need to update the object, using the <- assignment:

gapminder <- gapminder %>% 
  mutate(income_total = population*income_per_person)