Built for the SESA x National Bank Hackathon by:

• Jordan Yang
• Larry Wang
• Emily Hyuhn
• Sidharth Sajan

This project helps traders upload and analyze trade history to identify behavioral bias and improve risk discipline.
It includes bias detection, portfolio optimization recommendations, and an AI coach experience (Laurent Ferreira) with:

• Persistent chat history
• Insights panel (bias summary, charts, heatmap, suggestions)
• A moving mascot trigger (mascot-walking.gif) that walks around the page border

• Auth with Supabase
• Upload CSV or add trades manually
• Dashboard with:
  • P/L index by stock
  • 6-factor hexagonal bias radar (overtrading, loss_aversion, revenge_trading, disposition_effect, anchoring, confirmation_bias)
  • Recent trade table + quick clear
• All Trades page:
  • Server-side pagination/search
  • Per-trade delete and clear-all
• Bias Analysis and Risk Profile generation
• Portfolio Analysis:
  • Allocation charts and concentration metrics
  • Optimization recommendations
  • "Ask Laurent" handoff to AI coach with prefilled portfolio prompt
• AI Coach:
  • Dedicated coach page
  • Global moving mascot launcher
  • Clear chat history
  • Context-aware responses using trade, bias, insights, and portfolio optimization context

• Frontend: React, Vite, TypeScript, Tailwind, shadcn/ui, Recharts
• Backend services: Supabase (Auth, Postgres, Edge Functions)
• Optional analysis API: FastAPI + Polars (backend/)

• Node.js 18+ and npm
• Supabase project (URL + anon key)
• Optional: Python 3.12+ and uv for the FastAPI service

Copy the template and fill values:

cp .env.example .env

Required:

• VITE_SUPABASE_URL
• VITE_SUPABASE_PUBLISHABLE_KEY

Optional:

• VITE_API_URL (FastAPI endpoint, defaults to http://localhost:8000 if used)
• VITE_AI_COACH_AVATAR_URL (coach profile image)
• VITE_AI_COACH_MASCOT_URL (moving launcher image, defaults to /mascot-walking.gif)

npm install
npm run dev

Production build:

npm run build
npm run preview

cd backend
uv sync
uv run uvicorn app.main:app --reload --host 0.0.0.0 --port 8000

API docs: http://localhost:8000/docs

The app calls Supabase Edge Function ai-coach.

If you deploy/update it, ensure:

1. Function is deployed.
2. Secret LOVABLE_API_KEY is set in Supabase project secrets.

Parser supports flexible headers. Minimum required columns are:

• timestamp/date
• action (buy/sell)
• asset/symbol

Additional useful columns:

• quantity
• entry_price
• exit_price
• pnl
• account_balance
• notes

• npm run dev - start local frontend
• npm run build - production build
• npm run preview - preview build
• npm run lint - lint project
• npm run test - run Vitest tests

• src/ - frontend app
• src/pages/ - main product pages
• src/components/ - UI + coach components
• src/lib/ - data/services/insight logic
• supabase/ - migrations + edge functions
• backend/ - optional FastAPI analysis API
• docs/DEPLOYMENT.md - deployment notes

Use Vercel for frontend and Supabase for backend services.
See docs/DEPLOYMENT.md for deployment steps.