EcoPilot — Frictionless Carbon Tracker for India 🌿

EcoPilot (EcoCal) is a smart, dynamic carbon tracking application custom-designed for urban and semi-urban Indian smartphone users (ages 18–40). By translating abstract carbon metrics into concrete equivalents (trees, petrol litres, and money saved), it makes sustainability interactive, positive, and friction-free.

---

🚀 Chosen Vertical & Persona

• Target Audience: Indian mobile users who care about sustainability but will not tolerate complex homework, guilt-tripping interfaces, or carbon jargon.
• Product Promise: Zero-friction carbon tracking. EcoPilot guesses emissions from baseline onboarding answers and daily check-ins, allowing users to verify via vision captures (meals, power bills, retail receipts) without active manual calculations.

---

🛠️ Architecture & Tech Stack

   ┌────────────────────────────────────────┐
   │         React Native Expo App          │ (Expo Web support enabled)
   └───────────────────┬────────────────────┘
                       │ HTTP REST
   ┌───────────────────▼────────────────────┐
   │         Express Node.js Server         │
   └───────────────────┬────────────────────┘
                       │ Knex DB Wrapper
   ┌───────────────────▼────────────────────┐
   │      SQLite / PostgreSQL Database      │
   └────────────────────────────────────────┘

1. Frontend (/frontend): Built with React Native (Expo SDK 56) with Expo Router file-based navigation and full Web-bundler compatibility so it runs visually in browsers.
2. Backend (/backend): Node.js Express REST API using Knex query builder.
3. Database: Configured to run on SQLite locally for zero-configuration, instant startup, while remaining fully compatible with PostgreSQL in production by setting the DATABASE_URL environment variable.

---

📊 Core Calculation Logic & Approach

1. Zero-Friction Smart Onboarding

EcoPilot creates a starting footprint baseline in under 3 minutes through a 6-step survey wizard:

• Housing & Appliances: Proportional electricity baseline constructed from housing type, family size, AC ownership, and heavy appliance checkboxes.
• Diet: Baseline meal footprint derived from diet type (Vegan, Vegetarian, Eggetarian, Non-Vegetarian) and weekly non-veg frequencies.
• Commute: Travel baseline based on primary mode (Walk, Two-wheeler, Car, Transit, Mixed) and daily round-trip distance.
• LPG: Cylinders consumed per month (defaults to family size approximation if user is unsure).
• Air Travel: Domestic and international annual flights.

2. Daily Check-ins vs. Vision Captures (Override Logic)

• Check-in Deltas: Daily check-ins apply coarse multipliers (e.g. +50% travel emissions if user commuted more today) to estimate daily baseline adjustments.
• Vision Capture Overrides: Precise camera uploads (Food items, Electricity bills, Shopping receipts) act as overrides rather than stacking. When a capture exists for a category on a given day, it replaces the check-in's delta estimate for that category.
• Reminder Prompts: The backend monitors the user's upload timeline and dynamically generates dashboard reminders (monthly bill scans, monthly LPG updates, quarterly air travel counts).

3. Translation equivalents (convertCo2 Utility)

To make metrics relatable, EcoPilot converts raw kg CO₂ emissions using verified conversion keys:

• Trees equivalent: CO2 kg / 21 (absorptive capacity of 1 mature tree per year).
• Fuel equivalent: CO2 kg / 2.3 (emissions of burning 1 Litre of petrol).
• Cost savings: Blends tariff reductions (1 kWh ≈ ₹8) and fuel cost replacements (1 L petrol ≈ ₹100).

---

📋 Assumptions Made

1. Grid Emission Factor: Utilizes a single national grid average of 0.71 kg CO₂/kWh for India (based on CEA baseline data).
2. LPG Combustion: Assumes standard domestic cylinders (14.2 kg LPG) emit 42.5 kg CO₂ per cylinder.
3. Flight Constants: Emits a flat 250 kg CO₂ per domestic flight and 1100 kg CO₂ per international flight.
4. Vehicular Averages: Commutes assume 0.14 kg CO₂/km for cars, 0.04 kg CO₂/km for two-wheelers, and 0.025 kg CO₂/km for public transit.
5. Vision AI Mocking: Incorporates a mock LLM vision service returning pre-coded draft objects with a simulated 1.5s network delay to facilitate testing without requiring live Claude/Gemini API keys.

---

🏃 Setup & Local Execution

Prerequisites

• Node.js (v18+)
• npm (v10+)

1. Run Backend Server

cd backend
npm install
npm run dev

The server will start on port 3000 and automatically run database migrations and seeds on startup.

2. Run Frontend Web App

cd frontend
npm install
npm run web

The Metro Bundler will launch the application in your browser on http://localhost:8081.

🧪 Demo Credentials

• Mobile Number: 9876543210
• Verification OTP: 123456