🔮 Mira: AI-Powered CSV Data Analyst

Upload any CSV, get instant AI-powered insights with beautiful visualizations

🚀 Live Demo • 📊 Try Analysis

---

📋 Table of Contents

• The Problem
• The Solution
• Amazon Kiro Integration
• Key Features
• Architecture
• Quick Start
• Impact & Scalability
• Hackathon Submission

---

💬 The Problem

Last year, I was helping my uncle with his restaurant. He kept complaining about the POS system dumping massive CSV files every week—thousands of rows of orders, payments, inventory.

He'd open it once, get overwhelmed, and close it. "I just need to know if we're doing better than last month," he'd say.

Business owners don't need spreadsheets. They need answers.

---

💡 The Solution

Mira bridges the gap between raw data and actual understanding.

Upload any CSV file through our web interface, and within minutes receive:

• 📊 Automatic data exploration - AI understands your data structure
• 📈 Beautiful visualizations - Charts generated with matplotlib
• 🔢 Key metrics & KPIs - Numbers that matter, extracted automatically
• 💡 Actionable insights - Recommendations based on your data

No Excel skills required. No SQL knowledge needed. Just upload and get answers.

---

🛠️ Amazon Kiro Integration

This entire project was built using Amazon Kiro as the primary development environment.

How Kiro Accelerated Development

Amazon Kiro was instrumental throughout the entire development lifecycle:

Phase	How Kiro Helped
Architecture Design	Designed multi-agent architecture with E2B sandbox integration
Code Generation	Generated E2B agent logic, API routes, and React components
Debugging	Fixed Gemini model compatibility, function calling format issues
UI Development	Built responsive landing page and CSV upload interface
Deployment	Configured Vercel deployment with environment variables

Kiro Usage Screenshots

Development Session	Code Generation	Deployment
Architecture planning & setup	Building E2B agent & APIs	Testing & production deployment

Key Implementations with Kiro

// E2B Agent - Core analysis engine built with Kiro
export async function runE2BAgent(input: E2BAgentInput): Promise<E2BAgentOutput> {
  // Kiro helped design the iterative analysis loop
  // with proper Gemini function calling and error handling
  const sandbox = await Sandbox.create();
  // Multi-round analysis with automatic chart generation
}

// Web API - Clean endpoint for CSV analysis
export async function POST(req: NextRequest) {
  // Kiro created the complete analyze endpoint
  // with file upload handling and E2B integration
}

---

✨ Key Features

Feature	Description
🔐 Secure Sandboxed Execution	All Python code runs in isolated E2B sandboxes - no access to production systems
🧠 Multi-Step AI Agent	Iterative analysis with automatic data exploration, SQL-like queries, and KPI extraction
📊 Auto Visualization	Minimum 3 matplotlib charts per analysis - histograms, bar charts, trends
🎯 Smart Insights	AI extracts meaningful patterns, anomalies, and actionable recommendations
⚡ Fast Processing	Analysis completes in 1-2 minutes for most datasets
🌐 Web-Based	No installation required - works in any browser

---

🏗️ Architecture

┌─────────────────────────────────────────────────────────────────┐
│                         WEB INTERFACE                           │
│  ┌─────────────────────────────────────────────────────────┐   │
│  │              Next.js Frontend (React)                    │   │
│  │         Landing Page  •  CSV Upload  •  Results          │   │
│  └─────────────────────────────────────────────────────────┘   │
└─────────────────────────────────────────────────────────────────┘
                              │
                              ▼
┌─────────────────────────────────────────────────────────────────┐
│                         API LAYER                               │
│  ┌─────────────────────────────────────────────────────────┐   │
│  │              /api/analyze (POST)                         │   │
│  │         Receives CSV • Triggers E2B Agent                │   │
│  └─────────────────────────────────────────────────────────┘   │
└─────────────────────────────────────────────────────────────────┘
                              │
                              ▼
┌─────────────────────────────────────────────────────────────────┐
│                    AI ANALYSIS ENGINE                           │
│  ┌─────────────────────────────────────────────────────────┐   │
│  │              E2B Code Interpreter Sandbox                │   │
│  │  ┌─────────┐  ┌─────────┐  ┌─────────────┐             │   │
│  │  │ Pandas  │  │ SQLite  │  │ Matplotlib  │             │   │
│  │  └─────────┘  └─────────┘  └─────────────┘             │   │
│  └─────────────────────────────────────────────────────────┘   │
│                              │                                  │
│  ┌─────────────────────────────────────────────────────────┐   │
│  │           Gemini 2.0 Flash (Orchestration)               │   │
│  │      Function Calling • Multi-Round Analysis             │   │
│  └─────────────────────────────────────────────────────────┘   │
└─────────────────────────────────────────────────────────────────┘
                              │
                              ▼
┌─────────────────────────────────────────────────────────────────┐
│                         OUTPUT                                  │
│       Summary • KPIs • Charts (Base64) • Recommendations        │
└─────────────────────────────────────────────────────────────────┘

Technology Stack

Component	Technology	Purpose
Development	Amazon Kiro	AI-assisted development
Frontend	Next.js 15 + React	Web interface
Styling	Tailwind CSS	Responsive design
AI Orchestration	Google Gemini 2.0 Flash	Agent reasoning & function calling
Code Execution	E2B Code Interpreter	Secure Python sandboxes
Data Processing	Pandas + SQLite	CSV analysis
Visualization	Matplotlib	Chart generation
Deployment	Vercel	Serverless hosting

---

🚀 Quick Start

Prerequisites

• Node.js 18+
• E2B API Key (e2b.dev - $100 free credits)
• Google Gemini API Key (ai.google.dev - free tier)

Installation

# Clone repository
git clone https://github.com/AWS-25/mira.git
cd mira

# Install dependencies
npm install

# Configure environment
cp env.example .env.local
# Edit .env.local with your API keys

# Start development server
npm run dev

Project Structure

mira/
├── app/
│   ├── api/
│   │   └── analyze/route.ts    # CSV analysis endpoint
│   ├── analyze/page.tsx        # Upload interface
│   ├── page.tsx                # Landing page
│   ├── layout.tsx              # Root layout
│   └── globals.css             # Tailwind styles
├── lib/
│   ├── e2b-agent.ts            # Core AI agent logic
│   └── types.ts                # TypeScript definitions
├── assets/                     # Kiro screenshots
├── env.example                 # Environment template
└── README.md                   # This file

---

🎯 Impact & Scalability

Real-World Impact

Metric	Value
Time Saved	30 min manual analysis → 2 min automated
Accessibility	No Excel/SQL knowledge required
Cost	Free tier covers ~50+ analyses/month
Reach	Works on any device with a browser

Target Users

• Small Business Owners - Restaurant owners, retail shops, service providers
• Non-Technical Managers - Sales managers, operations leads
• Students & Researchers - Quick data exploration
• Anyone with CSV files - No technical skills required

Scalability

• Serverless Architecture - Auto-scales with Vercel
• Isolated Sandboxes - E2B handles compute scaling independently
• Stateless Design - Easy horizontal scaling
• Edge Deployment - Fast global response times

---

📦 Hackathon Submission

✅ Submission Checklist

Requirement	Status	Details
Project Documentation	✅	Complete README with architecture, setup, and impact
Working Code	✅	Clean TypeScript codebase
Live Demo	✅	mira-aws.vercel.app
Amazon Kiro Usage	✅	Screenshots + detailed integration documentation

🏆 Scoring Highlights

Criteria	Implementation
🛠️ Tool Integration	Extensive Amazon Kiro usage - architecture, code generation, debugging, deployment
🌐 Technical Quality	Production-ready, well-architected with E2B sandboxes + Gemini AI
🧾 Documentation	Comprehensive README with setup instructions and architecture diagrams
💡 Innovation	AI-powered CSV analysis accessible to non-technical users
🎯 Impact	Solves real problem - makes data analysis accessible to everyone
📈 Scalability	Serverless architecture with isolated compute sandboxes

---

Built with ❤️ using Amazon Kiro for AWS Vibeathon 2025

Making data analysis accessible to everyone

🌟 Try the Live Demo