Summary
"Messy data"—no matter what definition you use—present a huge roadblock for people who work with data. This chapter focused on two of the most notorious and prolific culprits: missing data and data that has not been cleaned or audited for quality.
On the missing data side, you learned how to visualize missing data patterns, and how to recognize different types of missing data. You saw a few unprincipled ways of tackling the problem, and learned why they were suboptimal solutions. Multiple imputation, so you learned, addresses the shortcomings of these approaches and, through its usage of several imputed data sets, correctly communicates our uncertainty surrounding the imputed values.
On unsanitized data, we saw that the, perhaps, optimal solution (visually auditing the data) was untenable for moderately sized datasets or larger. We discovered that the grammar of the package assertr provides a mechanism to offload this auditing process to R. You now have a few assertr checking "recipes...
