Add a changelog and populate it with draft 1.1.0 changes#85
Conversation
| data, you will need to use these new columns to handle changes to | ||
| agency numbers. See this vignette for a demonstration of how to | ||
| use the new columns to construct a crosswalk for agency numbers | ||
| across years: _TK: Link to vignette_ |
There was a problem hiding this comment.
Unless we change the .RMD file name, the link should ultimately be https://ccao-data.github.io/ptaxsim/articles/agencies.html
| - **How this change affects you**: You should read the latest version of the | ||
| vignette if you use PTAXSIM for TIF counterfactuals. | ||
| - **Added a new vignette to demonstrate the correct way to handle agency | ||
| numbers that changed in 2024 (_TK: Link to vignette_)** |
There was a problem hiding this comment.
Same as above: I'm waiting until #84 lands to add this link.
kyrasturgill
left a comment
There was a problem hiding this comment.
Thanks so much for documenting all of this! I just have some minor edits and suggested added context. Let me know if you want to talk more through how to frame the agency number change.
| consisted of three digits, and funds with the same `fund_num` in different | ||
| agencies always referred to the same type of fund. In 2024, the Clerk changed | ||
| their fund numbers so that they consist of six digits, and they are no longer | ||
| guaranteed to refer to the same type of fund across agencies. The new types | ||
| of funds that this change supports are primarily bond series. |
There was a problem hiding this comment.
| consisted of three digits, and funds with the same `fund_num` in different | |
| agencies always referred to the same type of fund. In 2024, the Clerk changed | |
| their fund numbers so that they consist of six digits, and they are no longer | |
| guaranteed to refer to the same type of fund across agencies. The new types | |
| of funds that this change supports are primarily bond series. | |
| consisted of three digits, and funds with the same `fund_num` in different | |
| agencies always had the same `fund_name`. In 2024, the Clerk changed | |
| their fund numbers so that they consist of six digits, and they are no longer | |
| guaranteed to share the same `fund_name` across agencies. The new types | |
| of funds that this change supports are primarily bond series. |
There was a problem hiding this comment.
Clarifying question before I accept this change: Were the names of funds with the same number indeed guaranteed to be identical prior to 2024? The line on the main branch that calculates fund_name in agency_fund_info makes it seem like they are perhaps similar, but not identical:
ptaxsim/data-raw/agency/agency.R
Line 154 in eb1e3e8
There was a problem hiding this comment.
My understanding was that this was done because of inconsistencies with case and spacing across years - I think that's ultimately why the table did not originally include agency_num as a key.
There was a problem hiding this comment.
Gotcha, thanks! I implemented your suggested change in fc6ecf7. I think an additional sentence like this might help clarify the situation even further:
In 2024, the Clerk changed their fund numbers so that they consist of six digits,
and they are no longer guaranteed to share the same name across agencies.
+ They also added many new six-digit funds that were previously aggregated into
+ larger three-digit funds. For these new six-digit funds, the first three digits of the
+ fund number always matches the three-digit number for the aggregate fund that
+ existed prior to 2024.Does this seem correct (and helpful) to you @kyrasturgill?
There was a problem hiding this comment.
Yes this makes sense! I was also thinking, should we clarify that for fund_num in years to 2024, we added "000" to make the format consistent?
Co-authored-by: Kyra Sturgill <49281603+kyrasturgill@users.noreply.github.com>
…` to match "Background" section
… in `NEWS.md` to match "Background" section
|
@kyrasturgill Thanks for the helpful review! Do you mind giving this another look? All of your comments should be resolved, except for #85 (comment), which I can resolve once we're on the same page. |
|
Dang yall did some work on this. |
|
It did some work on us, too 🙃 |
kyrasturgill
left a comment
There was a problem hiding this comment.
This all looks great to me! I really like the framing of "agencies that changed to funds" in 2024 - I think that is clearest way to explain it. I appreciate added context about fund_num and I defer to you if we want to add more explanation about our change to fund_num for years prior to 2024, but wanted to technically approve so it's ready to merge when you think it's ready.
|
@kyrasturgill I'm going to go ahead and merge this in, even though the agency crosswalk design is not fully settled yet (#88)! If we end up changing the crosswalk, we can tweak this changelog as part of that PR. |
This PR adds a new file
NEWS.mdthat we can use to document changes to the code and data over time. We populate it with draft changes for version 1.1.0, representing the necessary code changes powering the 2024 data release.NEWS.mdis a tidyverse standard for publishing release notes about R packages. Thepkgdownpackage that we use to build our docs has first-class support for it, and will automatically build it into our deployed docs.Here's a quick screenshot showing how the changelog renders on the
pkgdowndocs page:If you would like to render the doc in this format for review, you can run the following code in the RStudio console to launch a preview:
Closes #82.