Self-assessment: Exploratory data analysis with R

With this self-assessment you check whether you master the basics that exploratory data analysis starts with. If you have worked through the Whole game part of R for Data Science, or the paths Getting started with R and Data transformation with R and dplyr, most tasks should come easily. Try each one on your own first; you can open the solution if you get stuck or want to compare.

In every box the table rewe is ready, with a good 9,400 products from the REWE online shop: name, category, origin, price, organic label and nutrition values per 100 g. dplyr and ggplot2 are loaded. On the first click, loading takes a few seconds, and again for the first chart. Categories, countries and product names are in German.

Part 1: Overview

Task 1. Get an overview of rewe: which columns are there, what type do they have, and what do the values look like?

rewe

glimpse(rewe)

Task 2. The table has 9,438 rows. How many different products does it contain? Each product has its own product_id.

rewe

rewe |>
  distinct(product_id) |>
  nrow()

Part 2: Transforming

Task 3. Show the ten most expensive products, with name and price only.

rewe

rewe |>
  arrange(desc(price)) |>
  select(product_name, price) |>
  head(10)

Task 4. Which countries do the products come from? Show each country with the number of its products, the most frequent first.

rewe

rewe |>
  count(country_of_origin, sort = TRUE)

Task 5. Calculate a new column fat_unsat_in_gram, the fat content without the saturated fatty acids, and show it with name, fat and saturated fat.

rewe

rewe |>
  mutate(fat_unsat_in_gram = fat_in_gram - fat_sat_in_gram) |>
  select(product_name, fat_in_gram, fat_sat_in_gram, fat_unsat_in_gram)

Task 6. Which products have more than 99 g of fat per 100 g? Show only name and fat.

rewe

rewe |>
  filter(fat_in_gram > 99) |>
  select(product_name, fat_in_gram)

Task 7. Keep only the columns that hold numbers.

rewe

rewe |>
  select(where(is.numeric))

Task 8. product_id and gtin are numbers nobody calculates with. Convert both to text.

rewe

rewe |>
  mutate(
    product_id = as.character(product_id),
    gtin = as.character(gtin)
  ) |>
  select(product_id, gtin, product_name)

Part 3: Visualising

Task 9. Show in a bar chart how many products there are in each category.

rewe |>
  ggplot() +
  aes()

rewe |>
  ggplot() +
  aes(y = product_category) +
  geom_bar()

Task 10. Show in a histogram how the energy content in kcal is distributed across all products.

rewe |>
  ggplot() +
  aes()

rewe |>
  ggplot() +
  aes(x = energy_in_kcal) +
  geom_histogram(binwidth = 50)

Task 11. Compare the fat content of the categories with box plots.

rewe |>
  ggplot() +
  aes()

rewe |>
  ggplot() +
  aes(x = fat_in_gram, y = product_category) +
  geom_boxplot()

Task 12. Show with stacked bars what share of the products in each category is organic.

rewe |>
  ggplot() +
  aes()

rewe |>
  ggplot() +
  aes(y = product_category, fill = bio) +
  geom_bar(position = "fill")

Part 4: Looking closely

Task 13. A food has at most about 900 kcal per 100 g; not even pure fat has more. Find the products that have more, and show name, energy and fat. What has probably happened there?

rewe

rewe |>
  filter(energy_in_kcal > 900) |>
  select(product_name, energy_in_kcal, fat_in_gram)

Where to go next

If you managed almost everything without the solutions, you are well prepared. Where you got stuck, these help:

• Tasks 1 to 8: the path Data transformation with R and dplyr and the Transform part of R for Data Science.
• Tasks 9 to 12: the Visualize part of R for Data Science and, for the question of which chart fits, Fundamentals of Data Visualization.
• Task 13: there is no recipe here, only the habit of distrusting every result at first.