Snowmass Pod Recommendation MVP

Production-style MVP for recommending Snowmass ski pods using strict user constraints and Open-Meteo hourly forecasts.

For the simplest local launch on this computer, start with START_HERE.md.

Project Structure

• backend/: FastAPI API, terrain dataset, scoring logic, tests.
• frontend/: Static HTML/CSS/JS single-page UI.

Run locally

This project has two small parts:

• The backend is the brain. It picks the best Snowmass pod based on your settings and weather.
• The frontend is the page you open in your browser.

On Windows:

1. Install Python 3.11 from python.org.
2. Open PowerShell in this folder.
3. Install backend packages:

python -m pip install -r .\backend\requirements.txt

4. Start the backend:

python -m uvicorn app.main:app --host 127.0.0.1 --port 8000 --app-dir .\backend

5. In a second PowerShell window, start the frontend:

node .\frontend\server.js

6. Open http://127.0.0.1:3000
7. Quick checks:
   • Backend health: http://127.0.0.1:8000/health
   • Frontend: http://127.0.0.1:3000

Quickstart

1) Setup backend

make setup-backend

2) Run backend

make run-backend

Backend URL: http://127.0.0.1:8000

3) Run tests

make test-backend

4) Open frontend

Open frontend/index.html in your browser. The page calls http://127.0.0.1:8000/recommend.

API

• GET /health -> {"ok": true}
• POST /recommend -> top 3 pod recommendations (with best 2-hour window, reasons, and excluded pods)
• GET /live/snowmass -> live runs and lifts, with each mapped run including official pod metadata and a canonical difficulty label when the trail can be confidently linked

Recommendation Design

1. Apply strict hard constraints first.
2. Pull forecast hours from Open-Meteo.
3. Score each pod per hour and select best contiguous 2-hour window.
4. Return top pods with explainability bullets (2-5).
5. If weather unavailable, use neutral assumptions and reduce confidence.

Future Work

• Resort operations feed integration (lifts, closures, grooming reports).
• Better crowd modeling and dynamic congestion penalties.
• Run-level route suggestions beyond pod-level.
• Feedback-driven personalization from user labels.