Guided tour
Cowork Dashboard. One set of numbers three offices can trust.
An operations dashboard for our three offices. Weekly board exports feed most tabs, and the Cash Flow tab reads the Monday API every morning at 6. Nine tabs, and every metric is defined once. Each card counts the rows of its own drilldown, and a test checks it.
Walk through it, six steps
Four files, once a week
Ops exports the Leads board and the three patient boards to one Drive folder. That is the entire data engineering ask of the team, which is why it actually happens every week.
- CRM leads board export
- patient board, clinic A
- patient board, clinic B
- patient board, clinic C
The problem.
Leadership needed one view of the funnel, leads to scheduled to intakes to consults to accepted treatment, across Rochester, Buffalo and Syracuse.
- 31.2% to 5.8% after breaking surname ties on first name
- 19 minutes a finance request in the Feature requests tab, Oct 28
- 9 tabs every metric defined once
- Excel diff a harness checks metrics against ground truth
How it fits together.
Ingest
- Weekly Excel exports into one Drive folder
- A private service checks it every five minutes
- Header-name-based parser, order-proof
Model
- Each card counts its own drilldown rows
- Lead-patient join via CRM connect column
- Every metric defined once, for every view
Present
- Cloud Run web app behind Google sign in
- 9 tabs including Cash Flow and Feature requests
- Chart.js with prior-year overlays
Trust
- Data quality panel with fix instructions
- Python harness diffs metrics vs raw Excel
Where it landed.
Leadership reads one set of numbers that agree with themselves and with the raw exports, and the weekly ritual is light enough that it has survived contact with reality. The interesting engineering decision was restraint: swapping a brittle live-API design for a verifiable weekly pipeline raised trust more than any feature could. Clinic names and all figures shown are synthetic stand-ins.