🌿 EcoSync: Making Sustainability Second Nature

---

🚀 1. EcoSync: Making Sustainability Second Nature

The Hook

Why do traditional carbon footprint trackers fail? Because they force users to act like environmental accountants.

Existing carbon calculators suffer from two fatal flaws:

1. Manual Data Entry (High Friction): They ask users to input utility bills, odometer readings, and exact grams of waste. This leads to immediate churn.
2. Abstract Metrics (Climate Anxiety): Showing a user that they emitted "15.4 kg of CO₂" produces anxiety rather than action. To most people, a kilogram of gas is invisible and meaningless.

EcoSync is built to solve this. It redefines personal climate tech by moving from manual calculation to passive observation, and from abstract guilt to relatable, gamified agency.

┌────────────────────────────────────────────────────────────────────────┐
│  OLD WAY:                                                              │
│  "Enter your monthly gas bill in therms and electricity bills in kWh"   │
│  ❌ Immediate User Churn & Cognitive Fatigue                           │
├────────────────────────────────────────────────────────────────────────┤
│  ECOSYNC WAY:                                                          │
│  "Connect Google Fit & Plaid. We'll handle the rest."                  │
│  🚀 Zero-Friction Tracking & Real-World, Intuitive Actions             │
└────────────────────────────────────────────────────────────────────────┘

EcoSync introduces:

• Zero-Friction Ingestion: Automated scraping of travel and financial telemetry.
• Relatable Metrics: Translating invisible CO₂ emissions into standard everyday equivalents (e.g., the Cheeseburger Index).
• Hyper-Personalized AI Nudges: Context-aware, small behavioral swaps that require minimal effort but yield compounding environmental benefits.

---

🏛️ 2. The Core Architecture

EcoSync is composed of three interconnected pillars designed to deliver a premium user experience on top of a transparent, persistent database and a scientific calculation engine:

graph TD
    subgraph Client [Client - React Frontend]
        A[Onboarding Screen] -->|Lifestyle Selections| B[Carbon Dial Preview]
        C[Contextual Dashboard] -->|Cheeseburger Index & Status Badge| D[Daily Impact Gauge]
        C -->|Live Slider Adjustments| E[Calculator Sliders]
        C -->|Checklist Toggle| F[Actions Checklist]
        C -->|Visual Trend| G[Recharts Area Graph]
    end

    subgraph Server [Server - FastAPI Backend]
        H[onboarding.py] -->|Saves Profile| I[(SQLite DB: ecosync.db)]
        J[footprint.py] -->|Calculates & Updates Trend| I
        K[insights.py] -->|Generates via Gemini LLM| L[Insights Service]
        M[actions.py] -->|Updates Points & Challenges| I
    end

    subgraph Telemetry [Production Integrations]
        N[Plaid API] -.->|Financial Scrapes| H
        O[Google Fit API] -.->|Mobility Signals| H
    end

    A -->|POST /api/v1/onboard| H
    C -->|GET /api/v1/footprint/daily| J
    E -->|POST /api/v1/footprint/calculate| J
    F -->|POST /api/v1/actions/complete| M
    L -->|Real-time insights| K

Loading

🧠 Pillar 1: The AI Onboarding Coach (LLM Prompt)

To bypass forms, the onboarding flow is powered by a strict, conversational LLM prompt that extracts a user’s baseline using 3 simple lifestyle buckets:

1. Diet: Vegan, Vegetarian, Flexitarian, or Meat-Heavy.
2. Commute: Driving alone, Public Transit, Two-Wheeler, or Walking/Cycling.
3. Housing: Independent House, Apartment, or Shared/PG.

The onboarding system prompt processes conversational inputs, maps them to standard lifestyle classes, and inserts the profile directly into SQLite.

Note

The LLM System Prompt

System_Prompt: |
  You are the EcoSync Onboarding Coach, an empathetic, conversational AI agent designed to establish a user's carbon baseline.
  Your goal is to estimate the user's annual carbon footprint across 3 buckets (Diet, Commute, Housing) without asking for raw numbers or causing climate anxiety.

  ### Rules of Engagement
  1. Ask only ONE question at a time to prevent cognitive overload.
  2. Speak in a non-judgmental, encouraging, and clear tone.
  3. Convert descriptive answers (e.g., "I share an apartment with two friends and eat meat on weekends") into the standard classification.

  ### Lifestyle Buckets
  - Diet: [meat_heavy, flexitarian, vegetarian, vegan]
  - Commute: [drive, transit, two_wheeler, walk]
  - Housing: [house, apartment, shared]

  ### Output Constraints
  Once all 3 buckets are identified, output a final JSON block:
  {
    "onboarding_complete": true,
    "extracted_entities": {
      "diet": "<category>",
      "commute": "<category>",
      "housing": "<category>"
    },
    "conversational_summary": "Congratulations! Your baseline has been set. We estimated your impact at..."
  }

---

⚙️ Pillar 2: The Zero-Friction Engine (FastAPI + SQLite Backend)

The backend is built using Python FastAPI and is backed by a local SQLite database (ecosync.db) to ensure persistent tracking and point updates.

• API Configurations (.env): An optional GEMINI_API_KEY can be provided in the .env file. When present, the backend makes callouts to Gemini 1.5 Flash to generate hyper-personalized eco-insights. If no key is set, the system falls back to a smart, rule-based recommendation engine.
• Scientific Foundations: Calculations are backed by real-world emissions data from trusted sources:
  • EPA (US Environmental Protection Agency): Mobile combustion and fuel emission factors.
  • IPCC (Intergovernmental Panel on Climate Change): Global diet lifecycle emission factors.
  • Central Electricity Authority (CEA, India): Grids emission intensity factors (e.g., India national average ~0.82 kg CO₂/kWh, with regional grid factors to account for local coal vs. renewable ratios).

Regional Grid Factors & Emission Factors Table

Region	Core Cities Covered	CEA Grid Intensity (kg CO₂/kWh)	Transport Factor (kg CO₂/km)	Baseline Average Footprint (kg CO₂/yr)
Western India	Mumbai, Pune, Maharashtra	0.86	0.19	1,900 kg
Northern India	Delhi, Noida, Gurgaon	0.90	0.22	2,100 kg
Southern India (IT)	Bengaluru, Hyderabad	0.78	0.18	1,850 kg
Southern India (Coastal)	Chennai, Kochi, Kerala	0.72	0.17	1,750 kg
Eastern India	Kolkata	0.88	0.20	2,000 kg
India Fallback	All other cities / national average	0.82	0.21	2,000 kg
Global Fallback	International locations	0.49	0.17	4,700 kg

---

📊 Pillar 3: The Contextual Dashboard (React Frontend)

Built in React 18 with Vite and Tailwind CSS, the UI features a premium dark-mode aesthetic with vibrant eco-neon highlights.

• Interactive Tabs: Features a 3-tab layout:
  • Dashboard: Shows overall stats, the Recharts breakdown pie chart, Recharts history area graph, status badges, and AI insights.
  • Calculator: Adjust sliders for weekly commute distance, annual flights, monthly electricity, and diet type to instantly update the profile in the database and recalculate total footprint in real-time.
  • Actions: A checklist of habits to earn points, linked directly to progress bars for active community challenges.
• Cheeseburger Index: Rather than displaying abstract carbon figures in isolation, the dashboard converts CO₂ weight into the universally understood unit: Cheeseburgers (1 burger = 6.6 kg CO₂e baseline).
• Offline Resiliency: If the backend API is offline, the React app switches to local cached calculation state so the dashboard still renders while live connectors are unavailable.

---

📸 3. Visual Demonstration & UI Showcase

Judges can preview the high-fidelity dark-mode application interfaces below:

Onboarding Flow Wizard

A conversational onboarding form that live-estimates baseline carbon footprint before storing the profile in SQLite.

User Impact Dashboard

Features the live emissions breakdown dial, the Cheeseburger Index converter, monthly area trends, active community challenges, and Gemini AI-powered recommendations.

---

✨ 4. Key Features

• Automated Data Ingestion: EcoSync now exposes real connector readiness and a normalized /api/v1/footprint/sync endpoint for provider payloads from Google Fit, Plaid transactions, utility meters, and custom backend jobs.
• The Cheeseburger Index: By translating invisible emissions into food units, the user can understand their impact immediately (e.g., "My commute today was equivalent to 2.2 cheeseburgers").
• AI-Powered Smart Swaps: The backend sorts emissions categories to isolate the highest impact source and presents clear, achievable actions:

  Smart Swap Example: Swapping one beef meal this week for lentils cuts emissions by 3.2 kg CO₂ (equal to saving 1.3 cheeseburgers).

• Gamified Retention: Interactive checklists enable users to complete daily actions, accumulating points and driving progress on community challenges (like the "Meatless Monday streak" or "Zero-drive week").

---

🛠️ 5. Technical MVP Setup

Follow these instructions to run the EcoSync MVP locally.

⚡ Quick Start (Two Commands)

To run the application locally, open two terminal windows and execute:

Terminal 1 (Backend):

cd backend && pip install -r requirements.txt && uvicorn main:app --reload

• API Docs URL: http://localhost:8000/docs (Interactive Swagger UI)

Terminal 2 (Frontend):

cd frontend && npm install && npm run dev

• Dashboard URL: http://localhost:5173

---

💻 Detailed Setup Instructions

Prerequisites

• Python 3.10+
• Node.js 18+ & npm

Step 1: Start the Backend (FastAPI)

1. Open a terminal and navigate to the backend/ folder:

   cd backend

2. Create and activate a virtual environment:

   • macOS / Linux:

     python3 -m venv .venv
     source .venv/bin/activate

   • Windows (PowerShell):

     python -m venv .venv
     .venv\Scripts\Activate.ps1

   • Windows (CMD):

     python -m venv .venv
     .venv\Scripts\activate.bat

3. Install the required dependencies:

   pip install -r requirements.txt

4. Configure your .env variables: Copy the .env settings to enable the LLM functionality (add your key to GEMINI_API_KEY to test real LLM nudges).

   # ecosync.db will be created automatically in the backend folder
   GEMINI_API_KEY=YOUR_GEMINI_API_KEY

5. Start the FastAPI development server:

   uvicorn main:app --reload --port 8000

Step 2: Start the Frontend (React + Vite)

1. Open a new terminal window and navigate to the frontend/ folder:

   cd frontend

2. Install the frontend dependencies:

   npm install

3. Start the Vite development server:

   npm run dev

---

🐳 Step 3: Run as a Unified Docker Container (Optional)

EcoSync is designed to compile the frontend and backend into a single container for seamless deployment.

1. Build the Docker image from the root directory:

   docker build -t ecosync .

2. Run the container locally:

   docker run -p 8080:8080 ecosync

3. Open http://localhost:8080 to view the application.

---

🏆 6. Why EcoSync Wins (The Prompt Wars Edge)

EcoSync stands out in the Prompt Wars Challenge because it pairs advanced behavioral psychology with a robust technical foundation:

• Progressive Onboarding: Instead of bombarding the user with complex questionnaires on sign-up, the conversational AI coach extracts lifestyle vectors gently. By the time the user enters the app, their customized baseline is already calculated.
• The Guilt-Free UX: Most green apps trigger "climate guilt". EcoSync uses gamification, individual challenges, and positive reinforcement to build habit loops. (Note: Community features and cross-user streaks are planned for future releases).
• Privacy-First Telemetry: The application separates connector credentials from the UI. Provider keys and OAuth secrets stay in runtime environment variables, while the frontend only sees readiness state and normalized carbon inputs.
• Transparent Science: Unlike black-box carbon estimators, EcoSync links regional variables directly to international standards (IPCC/EPA) and local grid configurations (India Central Electricity Authority), with citations integrated directly into the user interface so you always know where your numbers come from.

---

Built with ❤️ for the Prompt Wars Challenge.