Chapter Eight
Biohealing Stream Quiz

Built withPython · FastAPI · App Engine · GitHub Actions
The Problem
Biohealing Stream visitors knew they wanted wellness support but had no way to tell which of 20 different Living Light protocols actually fit them. It needed to feel premium and private, finish in under a minute, embed cleanly on marketing pages without exposing the scoring logic, and give the team more than a submission count: real visibility into where people drop off and which protocols actually win.
What I Built
- Fifteen-question protocol matcher scoring across 20 distinct protocols (10 mind/spirit, 10 body-system), with the scoring map kept entirely server-side so it can never leak through the frontend.
- Weighted primary-focus scoring: question 1 asks for the user's single biggest concern and counts for 3x the others, a fix I made after testing found that three unrelated "experienced meditator" answers could silently outvote someone's actual stated concern and send them to "No Protocol" instead of the thing they asked about.
- A full automated test harness verifying all 20 protocols resolve correctly plus the edge cases (ties, multi-select body systems, the primary-focus override), run before any scoring change ships.
- Iframe-embeddable with full CORS support across Biohealing Stream, Biophotons.info, and any partner landing page.
- Session and drop-off tracking: every question advance fires a beacon, building an exact per-question funnel instead of just a completion count.
- Garden analytics dashboard, password-protected: protocol win rates, completion timing, device and traffic-source breakdowns, per-question answer distributions, and period-over-period comparisons, all filterable by date range.
- Branded result pages with protocol-specific artwork and anatomical diagrams for the body-system protocols, not just a text recommendation.
- Serverless deployment on Google App Engine (FastAPI, Python 3.11) with GitHub Actions auto-deploying on push, so there's no server to maintain.
How It Works
Each answer scores one or more of 20 protocols, with the user's stated primary focus in question 1 weighted 3x so it can't be quietly overridden by unrelated passive answers. Session tracking opens at quiz start, and a beacon fires on every question advance to build the drop-off funnel. Completed submissions persist to Firestore with full audit trails. The Garden dashboard surfaces protocol wins, funnel drop-off, and traffic source breakdowns, filterable and stamped with quiz version so a scoring change never corrupts historical comparisons.
Results
Live and embedded across the Biohealing ecosystem: a sub-one-minute assessment with zero server maintenance, CORS-safe iframe embeds, and real operator visibility into conversion funnels and protocol performance. Caught and fixed a real scoring bug, the primary-focus override, through actual end-to-end testing before it shipped to production.



At a Glance
- Assessment Time< 1 Min
- Protocols matched20
- Scoring Logic100% Server-side
- Deploy CI/CDGitHub Actions