Introduction to missing data
Dealing With Missing Data in R
Nicholas Tierney
Statistician
Introduction
The best thing to do with missing data is to not have any
--Gertrude Mary Cox
Working with real-world data = working with missing data
Missing data can have unexpected effects on your analysis
Bad imputation can lead to poor estimates and decisions.
What will you learn
- What missing values are
- How to find missing data
- How to wrangle and tidy missing data
- Explore why is data missing
- Impute missing values
Assumed knowledge
- Basic to intermediate experience with R
- Experience creating plots using ggplot2
- Experience using dplyr to manipulate and rearrange data
- Experience fitting linear models in R
What are missing values?
Missing values are values that should have been recorded but were not.
NA = Not Available.
How do I check if I have missing values?
x <- c(1, NA, 3, NA, NA, 5)any_na(x)
TRUE
are_na(x)
FALSE TRUE FALSE TRUE TRUE FALSE
n_miss(x)
3
prop_miss(x)
0.5
Working with missing data
NA + anything = NA
heights
Sophie Dan Fred
165 177 NA
sum(heights)
NA
Missing data gotchas
NaN: Not a Number.
any_na(NaN)
TRUE
any_na(NULL)
FALSE
any_na(Inf)
FALSE
Missing data gotchas (2)
NA | TRUE
TRUE
NA | FALSE
NA
NA + NaN
NA
NaN + NA
NaN
Let's practice!
Dealing With Missing Data in R
