How specific should `pull_{dataset}` functions be?

[dylanhmorris]

Right now we have pull_nhsn(). In #166, I am planning to deprecate it now that:

• we have more generic functions for interacting with data.cdc.gova
• but those do not expect the API endpoint as a URL the way pull_nhsn does.

Question 1: Provide a pull_data_cdc_gov_dataset() function?

This function would be similar to pull_nhsn(), providing some arguments for common filtering (e.g. start_date and end_date, location).

Not all datasets on data.cdc.gov will have a date column, a location column, etc., but we can design the function to fail appropriately in those cases.

I lean yes on Q1

Question 2: Provide pull_{dataset} alias functions for specific datasets we use a lot?

e.g. pull_nhsn_hrd_prelim(), pull_nssp_prop_ed_visits(), etc. as an alias for pull_data_cdc_gov_dataset() pointed at the NHSN HRD preliminary dataset, the NSSP ed visit dataset, etc.

I lean weakly no on Q2, as pull_data_cdc_gov_dataset("nhsn_hrd_prelim") seems fine.

[dylanhmorris]

Thoughts @sbidari @damonbayer @SamuelBrand1 @AFg6K7h4fhy2?

[damonbayer]

I am in favor of 2, because I would like to be able to access datasets using autocomplete in my IDE or the package documentation and not have to search the web to find the name of the data I am looking for.

[sbidari]

pull_data_cdc_gov_dataset("nhsn_hrd_prelim") seems fine to me.

[dylanhmorris]

I'm only going to implement pull_data_cdc_gov_dataset(<name>) in this #166 but we can consider adding pull_<name>() aliases in a separate PR.

[dylanhmorris]

We seem to be doing just fine with pull_data_cdc_gov_dataset(<name>), so I am going to close this as decided for now.