Exercises - Programming Like a Data Scientist

Basics of R for Data Science

This set of exercises includes practicing and integrating various skills: working with dataframes and lists, applying conditional logic and iterative programming. Some of these exercises are pretty advanced for the purposes of the present course!

• Download this workspace and import it using the load() function.

• The dataframe df1 has all its columns stored as character, but many are mostly numeric, and should be treated as such. Your task is to determine which columns can reasonably be coerced to numeric. Use a combination of iterative and conditional programming to complete the following steps:

  • Iterate over all columns of df1 (using a for loop is recommended);
  • For each column, calculate the percentage of observations that remain valid (i.e., do not become NA) when coerced to numeric using as.numeric();
  • If the percentage exceeds 80%, coerce the column to numeric;
  • Keep track of the columns that were coerced to numeric;
  • After processing all columns, compute a correlation matrix that includes only the columns that were coerced to numeric (use the cor() function; set argument use="pairwise.complete" to avoid losing too many observation with listwise-deletion);
  • for fun: use the corrplot() function from the corrplot package to obtain a more colorful correlation matrix

• Extra for advanced users: The dataframe dfWide1 contains data for treated subjects at three times ("T0", "T1", and "T2") in wide format, but it needs to be reshaped to long format for data analysis. Your task is to reshape it to long format using the basic reshape() function (first of all, have a look at the documentation of reshape()).

• Extra for advanced users: Repeat the previous task using the convenient pivot_longer() function from the tidyr package.

• Extra for super advanced users: The dataframe dfWide2 is even more complex: as it contains two variables, Acc and RT, measured at three times. Reshape dfWide2 to long format, but keeping Acc and RT values in separate columns. Use any approach you like, but try to keep the code compact if possible.