Load libraries in data-generating scripts

[nmdefries]

Also do some light cleanup to data scripts to make sure they run and script, output file, and output data object names all match.

I've dropped the cancovid.R script since it depends on non-included raw data files, and is not exported/documented by the package. cancovid.R has been reconstructed based on code from https://github.com/zcaiElvis/Covid19Canada

[brookslogan]

Nice cleanup, looks good.

Only potential blocker: @dajmcdon do you want to keep cancovid somewhere?

[nmdefries]

Differences between old and new versions of the Canadian case rate dataset:

• The old one doesn't have data for New Brunswick
• I changed all province names to long form, for consistency (e.g. "British Columbia" instead of "BC", in the old data)
• More time values are available for each version in the new dataset. In the old script, the version dates got misaligned from the data, and each issue was actually older than the version date label and thus had fewer recent time_values available.
• Case rates are overall slightly different in the new version, also because of the version alignment issue.

None of these appears to be relevant to the epipredict vignette that this dataset is used in. @dajmcdon to confirm.

[nmdefries]

@dajmcdon reminder to look at the new version of the cancovid script

[dajmcdon]

@nmdefries Thanks for the reminder. A few minor things:

1. Some enhanced docs for cancovid.R
2. I think that the epidatr calls are different now, so we should probably update.