Add measles and EPICC, CRAN download badges, and fix the dashboard metrics

[gvegayon]

What this does

1. Two projects added to data/

Project	Repo	Notes
measles: Measles Epidemiological Models	UofUEpiBio/measles	Description, license, and authors taken from its DESCRIPTION. On CRAN.
EPICC: Epidemiological Cost Calculator	EpiForeSITE/epicc	Description assembled from the repo README and docs/stories.md.

Two things worth a second opinion from the maintainers:

• epicc has no license file, so license is .na.character. Worth adding one to the repo.
• Contacts are a guess from the commit history (measles: George; epicc: Edward Wang and George). Correct them if that is not who should be listed.

2. CRAN badges in the dashboard

New CRAN column with version, grand-total downloads, and downloads/month badges. Seven projects match today: epitraxr, epiworldR, linkeR, measles, pomp, quadkeyr, rbranding.

How a project is matched to a CRAN package:

1. The optional cran_name field in the project's YAML wins. Use it when the names differ (epiworld now declares cran_name: epiworldR, since its CRAN artifact lives in another repo), or set it to no to skip the lookup entirely.
2. Otherwise the repository name and the first word of tool_name are checked against the CRAN index, and a hit is only accepted when the package's CRAN metadata (URL or BugReports) points back at the same GitHub owner or repository.

Step 2 is what keeps a project from picking up the badges of an unrelated CRAN package that happens to share its name, and it still catches the cases where repo and package names have drifted apart (quadkeyr moved to rOpenSci, rbranding lives in branding-package).

3. Bugs fixed along the way

The numbers in the dashboard were wrong, in ways that were invisible because most of them were never printed:

• Contributors and open PRs used length() on a parsed data frame, which returns the column count. Every repository was recorded as having 20 contributors and 36 open PRs. Only the avatars were displayed, so this stayed hidden.
• Open issues came from open_issues_count, which GitHub defines as issues plus pull requests. PRs are now discounted, so the number agrees with the issues badge next to it.
• Commit totals came from the search API, which is heavily rate limited and returns 0 for repositories it has not indexed. epiworld was reported as having 0 commits; it has 658. All collection sizes now come from the last page number in the Link header: one request, exact count, no fallback chain.
• A failed request now reports N/A instead of 0.
• Table rows had one more cell than the header.
• commits_data <<- was writing to the global environment.

4. Presentation

• Projects are sorted by most recent commit, each with an activity marker: 🟢 within 90 days, 🟡 within a year, ⚪ older or unknown.
• A short summary above the table: how many projects are active / slow / dormant, how many are on CRAN, and the total open issues and PRs.

Notes for reviewers

• README.md, profiles.md, and data/software.xlsx are not in this PR. They are generated by the Build README workflow, which will regenerate them on merge. README.qmd now uses rbindlist(fill = TRUE) so the optional cran_name field does not break that render.
• No new R package dependencies. CRAN lookups use available.packages() (one call per run) plus crandb.r-pkg.org for the handful of candidate hits.

Suggestions not included here

• PyPI badges for the Python tools (epiworldpy), which the CRAN code is shaped to accommodate.
• vaccine-equity-model and multigroup-vaccine are duplicate entries for what looks like the same project, under the same name, with identical metrics. One of them should probably go.
• Fernandez-Lab-WSU/geocovid_app returns 404 to the API, so GeoCovid shows no data. It looks like the repo was moved, renamed, or made private.
• The data/*.yml files mostly lack a trailing newline, which makes the script emit ~40 "incomplete final line" warnings per run.

🤖 Generated with Claude Code

[gvegayon]

Follow-up pushed in 78b3d09:

• CRAN badges moved into the Repository column, on one line under the repository link. The separate CRAN column is gone.
• epiworldRcalibrate added (data/epiworldrcalibrate.yml) — repo is sima-njf/epiworldRcalibrate, classified as Parameter estimation, contact taken from the DESCRIPTION (Sima Najafzadehkhoei).
• rbranding was already listed, but pointed at EpiForeSITE/branding-package, which has since been renamed to EpiForeSITE/rbranding; its docs link moved with it. Both are updated, and the package name is now declared explicitly instead of inferred.

The dashboard now tracks 24 projects, 8 of them on CRAN.