Keep the local copies for now: repoint first, delete after the next publish

Splits this PR into the first of two phases, because deleting the files here
breaks the CURRENTLY PUBLISHED site rather than this repo.

This series publishes on a `publish*` tag, not on push to main, and was last
published on 2026-07-17. So merging a deletion does not refresh the live site
-- it leaves the already-published notebooks pointing at a file that no longer
exists. That is not theoretical: set 1 (mpd2020.xlsx) merged with its deletion
included, and the notebook currently served at intro.quantecon.org still
carries the old URL, which now returns 404. Anyone downloading or opening that
lecture in Colab today hits it. The rendered HTML pages are unaffected, since
their figures are baked at build time; the breakage is confined to readers who
actually run the code, which for this series is the audience that matters.

So the ordering rule this migration has been following was incomplete.
"Repoint every consumer together" protects the REPOS but says nothing about
the published site, which lags main until someone tags a release.

Two phases close the gap completely:

  1. this PR -- repoint the URLs, keep the files. Publish. The published
     notebooks now fetch from data-lectures while the old paths still resolve,
     so neither the old nor the new site can break.
  2. a follow-up PR -- delete the files, once nothing references them in the
     repo OR on the live site.

The cost is one extra PR per repoint set and a slightly slower orphan
cleanup. The benefit is that no reader-facing window exists at any point.

lecture-wasm's half needs no such split: it publishes on push to main, so it
self-heals on merge, and the copies it deletes are mirror-orphans that nothing
reads in either repo.

See QuantEcon/data-lectures PLAN.md, "Repoint rules".

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

Author: mmcky