Missing values

Question 1

generate the summary statistic abs visualizations of dataset

Answer 1

skim()

Question 2

function is useful for adding rows for the missing combinations of variables

Answer 2

complete()

Question 3

visualize missing data

Answer 3

vis_miss()

Question 4

There are three main types of missing values:

Answer 4

1. Missing Completely at Random (MCAR). “The dog eats homework”
2. Missing at Random (MAR). “The dog ate a particular student’s homework”
3. Missing not at Random (MNAR). “The dog only eats bad homework”

Question 5

command is useful to use alongside a
filter
to just get the complete rows. (missing values)

Answer 5

complete.cases()

Question 6

Mean value imputation

Answer 6

Replace any missing values with the mean of the available data for numeric variables.
Replace missing values with the modal (i.e. most common) category (level).
Very simple to implement.
Very crude – can distort structure of dataset

Question 7

function can do mean imputation

Answer 7

impute_mean()

Question 8

Nearest neighbour imputation

Answer 8

We do imputation based on records that are similar to the one with missing data. Can measure similarity (or rather dissimilarity) by calculating a distance between records. Could use Euclidean (straight line) distance.
Or some other criteria.