Opentra

Operational observability for Indian D2C brands.

What is included

• Next.js 14, TypeScript, TailwindCSS, Zustand, Recharts, Lucide frontend
• FastAPI backend skeleton with SQLAlchemy models, PostgreSQL, Redis/Celery wiring, Pandas upload preview
• Fuzzy column mapping flow with user confirmation
• Baseline-mode first upload handling
• Deterministic signal thresholds in code
• Campaign-level RTO spike signal as the core demo decision
• Realized ROAS and RTO calculation helpers that use delivered orders only

Run frontend

npm install
npm run dev:infra
npm run dev

Open http://localhost:3000.

Run backend

npm run dev:infra
cd apps/api
python3 -m venv .venv
.venv/bin/activate
pip install -r requirements.txt
npm run dev

API health check: http://localhost:8000/health.

Non-negotiable commerce rules

• RTO rate is calculated as returned_orders / delivered_orders * 100
• Realized ROAS is calculated as revenue_from_delivered_orders / ad_spend
• First upload creates a baseline snapshot and does not produce false monitoring inferences
• Signal detection is deterministic and threshold-based for MVP

Upload Flow

How It Works

Users can now upload Excel files directly without selecting a source type. The system automatically detects the data source based on sheet names:

• Shopify Orders: Sheets containing "shopify" or "order" in the name
• Meta Ads: Sheets containing "meta" or "ad" in the name
• Inventory: Sheets containing "inventory" in the name
• Creative Performance: Sheets containing "creative" in the name
• Customer Signals: Sheets containing "customer" in the name

Multi-Sheet Uploads

You can include multiple sheets in a single Excel file (e.g., Shopify Orders, Meta Ads, and Inventory all in one workbook). The system will process each sheet according to its name.