Update the variable dictionary to add human-readable transformations for values

[jeancochrane]

We apply some transformations to model features in https://github.com/ccao-data/pinval in order to make the values more readable, e.g. formatting other_tax_bill_rate as a percentage and formatting acs5_median_income as currency. It would be nice to encode these transformations in vars_dict so that we can use them more widely.

See Dan's original comment (ccao-data/homeval#18 (comment)) for more context.

Data dictionary update checklist

• Make your updates to the source data file and commit them in a branch.
• Run data-raw/<data_dict>.R. This will save the CSV as an R data file (data/<data_dict>.Rda). This file is what the unit tests and package run against.
• Update the docstring for the data dict in R/data.R. This is where help docs for the packaged data.frame are kept. It needs to be updated manually to reflect the changes to the raw data.
• Run devtools::test() to make sure all unit tests pass with the new dictionary.
• Run devtools::document() to generate new man files from the changes to R/data.R.
• Run devtools::check() to run a test package build.
• Run pkgdown::build_site() and navigate to the reference page to make sure the changes are reflected on the docs website.
• Commit changes made by the preceding commands, push them to GitHub, and request a review.