In production · Harborview Metro Health District
Dashboard or reportWastewater surveillance trend alerts
Result in one line: Epi team saw the winter flu rise nine days before case reports confirmed it
Watches the weekly wastewater sampling results for the metro's five treatment plants, flags statistically unusual rises for COVID, flu and RSV, and posts a short plain-language alert to the epi team's channel before the Monday briefing.
Published August 26, 2026
- Skills needed to set it up
- Analyst or data scientist
- Readiness
- Guided setup Needs customization Human review built in
- Data it touches
- Public data only De-identified data
- Who sees the output
- Internal staff
- License
- MIT
- Cost to stand up
- No new spend
- Portable to other platforms
- Partially — with rework
- Use case category
- Operations & logistics
- Reviews it went through
- Privacy review AI governance body
What do these mean?
- Analyst or data scientist
- Someone who works in Python, R or SQL.
- Guided setup
- A wizard or script walks you through installation.
- Needs customization
- Your team will need to adapt it before use.
- Human review built in
- A person checks the AI's output before it is used.
- Public data only
- Only data that is already public.
- De-identified data
- Personal identifiers removed before use.
- Internal staff
- Used only inside the organization.
- MIT
- Permissive; reuse with attribution.
- No new spend
- Built with licences, staff and infrastructure the organization already had.
- Partially — with rework
- Some pieces are vendor-specific and would need swapping.
- Operations & logistics
- Planning, dispatch, inventory, inspections and field work.
- AI governance body
- An internal AI review board or committee signed off.
We started posting wastewater trends by hand in 2024 and kept missing rises that were obvious in hindsight. The alerting version fits a seasonal baseline per plant and per target, flags a sustained two-sample rise above the baseline’s expected band, and drafts a three-sentence note an epidemiologist approves before it posts. The human approval step matters: about one flag in five is a sampling artifact, and the reviewer catches those in under a minute.
Setup for another jurisdiction means pointing the ingest at your lab export, setting plant service-area names, and two seasons of history for the baseline. The alert prompt is in the repo and is deliberately boring — it summarizes numbers the model computed, it never speculates about causes.
About
- Area of work
- Epidemiology & surveillance
- Data & informatics
- Review status
- Reviewed & approved
How it's built
- How AI is involved
- AI is part of the solution
- Types of AI
- Prediction & forecasting
- Classification & NLP
- AI tools & models
- Python
- Prophet
- Claude (API)
- Where it runs
- Microsoft Azure
Sharing & licensing
- Portability notes
- The trend model and alert text generation are platform-agnostic Python; the ingest step assumes our LIMS export format and would need a new adapter for a different lab system.
What it took
- Cost to keep running
- Under $10k/yr
- How it was bought
- Existing enterprise licence
- Who it affects
- Alerts are internal decision support for epidemiologists. Sampling covers the whole metro sewershed, so no neighborhood is singled out; the alert text never names a facility smaller than a treatment plant service area.
Data & access
- No PII/PHI in the shared material
- Yes
- Data sources
- Weekly wastewater sample results from the five metro treatment plants
- State respiratory dashboard (public) for corroboration
- Data-governance caveats
- Plant-level aggregate counts only — no case-level or personal data anywhere in the pipeline. Alerts are internal decision support for the epi team; the public state dashboard is used to corroborate trends, not fed by this tool.