Skip to content

Draft cohort analysis#344

Open
RichardJECooke wants to merge 88 commits into
mainfrom
draft-cohortAnalysis
Open

Draft cohort analysis#344
RichardJECooke wants to merge 88 commits into
mainfrom
draft-cohortAnalysis

Conversation

@RichardJECooke
Copy link
Copy Markdown

sixhobbits and others added 30 commits April 24, 2026 11:26
* Add agent-identity-walled-gardens blog post, publishing 2026-04-24

* Google Cloud 2026 dev keynote post

* Update agent-identity-walled-gardens.mdx

* Update agent-identity-walled-gardens.mdx

* Update agent-identity-walled-gardens.mdx
@sixhobbits sixhobbits force-pushed the draft-cohortAnalysis branch from 4317956 to e4a316b Compare April 30, 2026 14:42
@sixhobbits sixhobbits force-pushed the draft-cohortAnalysis branch from e4a316b to f521c5e Compare April 30, 2026 14:53
jamesdanielwhitford and others added 9 commits May 14, 2026 15:44
Dan reported the extract took ~13 minutes for ~20k users and was
unsure if it had stalled. Add a short note after the extract command
so readers expect the wait and know progress is being printed.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The note was added when the example repo defaulted to 20,000 mock
users (a ~13 minute extraction). The default has been reverted to
1,000, which completes in well under a minute, so the warning is
no longer warranted.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The previous "Charts created" line was incomplete — charts.go also
prints "charts-output.html written" and the actual ready signal
"Server listening at http://0.0.0.0:7777". The first line appears
before the server is up; a reader watching only for it would think
the program had hung after rendering.

Also surface charts-output.html, the standalone HTML the program
writes on startup. It's mentioned in the output but had no
explanation; without one, readers would wonder what the file is.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Two prose fixes in the cohort analysis article.

1. Abandonment chart description. The previous wording ("haven't
   logged in for more than one, two, six, or 12 months") naturally
   reads as cumulative thresholds, but the code (and the rendered
   image) use mutually exclusive buckets, with the last one
   open-ended. Cumulative would be mathematically impossible given
   the image: it shows ~560 users in the "12" column and only ~60
   in "1", but if 560 users were inactive for 12+ months they would
   all also count as inactive for 1+ month, so "1" would have to be
   at least 560. Rewrite to describe the actual exclusive bucketing
   and explain why "12+" dominates.

2. Appendix instruction "Adjust the URL, authorization key, and
   application Id in the code above" was nonsensical — those values
   live in .env, and no code appears above. Drop "in the code
   above".

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The accompanying repo no longer ships a tracked .env (it now ships
.env.example with placeholder values). Update both spots that
referenced .env to point at the new copy-and-edit flow.

The appendix spot also gets a "if you skipped the extraction step"
qualifier — a reader running the appendix after the main extraction
section already has a working .env and doesn't need to do this
again.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The accompanying repo no longer writes a static HTML copy on
startup; the server at localhost:7777 is the only output. Update
the expected terminal output to match and remove the now-stale
sentence about a standalone copy.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

9 participants