I'm a technical writer with 10 years of experience in docs-as-code (Markdown, Git, CI/CD) and the Diátaxis framework. I use SQL and Python to measure reader value, find what is missing, and know where to spend time improving documetation.
dochealth is a Python tool and Streamlit dashboard that measures the health of any docs-as-code repo. It reads git history and Markdown to provide a page-by-page report on the following:
- Freshness: days since last content change, median line age
- Readability: Flesch reading ease, word count, code density
- Structure: headings and internal links
- Ownership: authors and commit counts
The screenshot shows the output from a run over Docusaurus's own docs. It surfaces pages that look maintained but aren't: each was edited in the last four months, yet the median line is more than four and a half years old.
Live dashboard → · Case study → · Repo →
| Project | What I did |
|---|---|
| Shoreline | Migrated 200+ pages from Freshdesk to Docusaurus in under 12 months, restructuring the content along Diátaxis lines. Cut onboarding time from weeks to hours. |
| super.AI | Wrote ML data-platform docs from scratch: 100 pages, including an API reference covering 30 endpoints, video guides, concept pages, and getting started guides. |
| Adjust Help Center | Built the client-facing help center in Salesforce and the style standards behind it. |
| Adjust features | Documented 10 major and 50+ minor releases, plus UI copy. |
Docs
Workflow
Data
Reading philosophy, poetry, and novels · fingerstyle acoustic guitar in the John Fahey vein · running (one marathon so far)