E-Commerce BI Platform - Public Portfolio Copy

A production-level full-stack business intelligence platform for fashion e-commerce brand operations. The app ingests marketplace sales, inventory, returns, advertising, visibility, and sale-grade files, then turns them into executive KPIs, category decisions, replenishment recommendations, anomaly alerts, and downloadable business reports.

This repository is a public portfolio-safe version. Real credentials, raw business exports, generated reports, and production data are intentionally excluded.

Highlights

• Full-stack BI app with FastAPI, Next.js, Supabase Postgres, and Tailwind CSS.
• Secure Supabase Auth with role-based access for admin, management, analyst, MD, and viewer users.
• Daily upload center for sales, inventory, returns, ads, visibility, SKU mapping, and sale-grade files.
• Marketplace analytics across Myntra, Ajio, Nykaa, Flipkart, Amazon, and TataCliq.
• Executive dashboard with MTD KPIs, sales trends, channel mix, category mix, sales-vs-returns, and regional map views.
• Inventory dashboard with OOS, broken stock, DOI, replenishment priority, category/status matrix, and CSV downloads.
• Category engine for NOOS, Potential NOOS, Green, Yellow, Red, Dead, Winter, Dog styles, Discontinue, Watchlist, and launch/repeat statuses.
• Returns and sales reports with export safety for CSV/Excel formula injection protection.
• AI assistant integration with Claude first and OpenAI fallback.
• Anomaly alert system for style-level sales spikes, return spikes, overstock risk, and category inventory movement.
• Deployment-ready backend and frontend configs for Railway and Vercel.

Tech Stack

Layer	Technology
Backend	FastAPI, Pydantic, Pandas, Supabase Python client
Frontend	Next.js App Router, React, TypeScript, Tailwind CSS
Charts	Recharts, react-simple-maps
Auth	Supabase Auth plus custom role table
Database	Supabase Postgres with migrations and RLS policies
AI	Anthropic Claude with OpenAI fallback
Notifications	Twilio WhatsApp and Resend email integrations
Deployment	Docker, Railway, Vercel
Testing	Pytest, Next.js build validation

Architecture

flowchart LR
    Files["CSV/XLSX marketplace exports"] --> Upload["FastAPI upload routers"]
    Upload --> Ingestion["Pandas ingestion services"]
    Ingestion --> Supabase["Supabase Postgres"]
    Supabase --> Routers["FastAPI analytics APIs"]
    Routers --> Frontend["Next.js dashboard"]
    Frontend --> Users["Executives, ops, category, and inventory teams"]
    Routers --> Reports["CSV / Excel / PNG reports"]
    Routers --> Assistant["AI assistant"]

Loading

Screenshots

The screenshots below are rendered from the repo's public /demo route using sample data, so recruiters can view realistic dashboard output without private credentials or live business data.

Main Capabilities

Executive Analytics

• MTD sales, quantity, ASP, returns, OOS, broken stock, and inventory KPIs.
• Daily sales trend and sales-vs-returns comparison.
• Marketplace and channel contribution views.
• Category mix and India regional heat map.

Sales And Returns

• Channel-wise daily sales trends.
• Marketplace sales/return summaries.
• Top products by revenue, ROS, growth, and return percentage.
• Same-period growth comparisons.
• Return trend, high-return SKU analysis, and returns report downloads.

Inventory And Replenishment

• Current inventory health by style and status.
• DOI rules for OOS and zero-ROS stock.
• Category/status matrix with style counts and inventory sums.
• Replenishment priority planning with download support.

Category Engine

• Rule-based category assignment with sale-grade overrides.
• Potential NOOS queue based on ROS and grade eligibility.
• Sale-grade master upload and category rebuild endpoint.
• Cross-analysis of old sale grade vs new category.

Alerts And Reports

• Style-level anomaly alerts for sales spikes, return spikes, combined sales/return spikes, and overstock risk.
• Category-level movement alerts based on inventory depth and SKU count change.
• Filterable alert table and downloadable alert reports.
• Daily Sale & Return report, inventory report, returns report, replenishment report, and category report.

Repository Layout

backend/
  db/migrations/        Supabase schema and supporting tables
  routers/              FastAPI route modules
  services/             Ingestion, category, alert, report, and forecasting logic
  models/               Pydantic schemas
frontend/
  app/                  Next.js App Router pages and API proxy routes
  components/           Dashboard UI, charts, forms, and layout components
  lib/                  API clients, auth helpers, formatting, date range helpers
  public/               Non-sensitive UI assets and map topology
tests/                  Backend and business-rule regression tests

Local Setup

Prerequisites:

• Python 3.11
• Node.js 18 or later
• Supabase project

1. Clone the repo.

git clone https://github.com/Sandy7217/ecommerce-bi-platform.git
cd ecommerce-bi-platform

2. Create environment files from examples.

Copy-Item .env.example .env
Copy-Item frontend/.env.local.example frontend/.env.local

3. Fill in Supabase and optional integration credentials.

Backend .env:

SUPABASE_URL=https://your-project-id.supabase.co
SUPABASE_ANON_KEY=your-anon-public-key-here
SUPABASE_SERVICE_KEY=your-service-role-secret-key-here
LEAD_TIME_DAYS=45
ENVIRONMENT=development

Frontend frontend/.env.local:

NEXT_PUBLIC_SUPABASE_URL=https://your-project-id.supabase.co
NEXT_PUBLIC_SUPABASE_ANON_KEY=your-anon-public-key-here
NEXT_PUBLIC_API_URL=http://localhost:8001

4. Install dependencies.

python -m pip install -r requirements.txt
npm install

5. Apply database migrations in Supabase SQL Editor.

backend/db/migrations/001_initial.sql
backend/db/migrations/002_targets.sql

6. Start the backend and frontend.

python -m uvicorn backend.main:app --host 0.0.0.0 --port 8001
npm run dev -- --port 3001

Open http://localhost:3001.

Validation

Run the backend regression suite:

python -m pytest -q

Run the frontend production build:

npm run build

Public Data And Secrets Policy

This portfolio repo does not include:

• Raw marketplace reports
• Real Supabase keys or project secrets
• .env or frontend/.env.local
• Generated reports, logs, screenshots with live business data, or Vercel local project files
• Git history from the private working repository

Use your own Supabase project and sample data to run the app.

Engineering Notes For Recruiters

This project demonstrates:

• End-to-end BI system design from raw operational files to executive dashboards.
• Business-rule-heavy backend engineering with regression tests.
• Secure full-stack auth integration across server-rendered and client-rendered routes.
• Data ingestion with heterogeneous CSV/XLSX schemas.
• Dashboard UX for repeated operational use rather than marketing pages.
• Deployment hygiene for split frontend/backend hosting.

License

Portfolio demonstration project. Add a license before using this code commercially.