Evaluating imputations: Across many variables
So far, we have covered ways to look at individual variables or pairs of variables and their imputed values. However, sometimes you want to look at imputations for many variables. To do this, you need to perform some data munging and re-arranging. This lesson covers how to perform this data wrangling, which can get a little bit hairy when considering its usage in nabular
data. The function, shadow_long()
gets the data into the right shape for these kinds of visualizations.
This exercise is part of the course
Dealing With Missing Data in R
Exercise instructions
- Use the
shadow_long()
to gather the imputed dataocean_imp_mean
, focussing onhumidity
andair_temp_c
. - Print the data and inspect it.
- Explore the imputations in a histogram using
geom_histogram()
, placing the values on the x-axis, filling by their missingness and faceting byvariable
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Gather the imputed data
ocean_imp_mean_gather <- shadow_long(___,
___,
___))
# Inspect the data
___
# Explore the imputations in a histogram
ggplot(ocean_imp_mean_gather,
aes(x = value, fill = value_NA)) +
geom_histogram() +
facet_wrap(~variable)