☁️ InsightCloud

Turn CSV files into conversational, AI-driven insights — no coding required.

InsightCloud is a cloud-native SaaS platform where users upload structured datasets and ask questions in plain English. The AI responds with specific, data-backed answers drawn from the actual uploaded data — not generic responses.

Built with Python, Streamlit, Pandas, and AWS (S3, Bedrock, EC2), with a free local development mode using Groq.

---

How It Works

Upload CSV → Pandas analyzes it → You ask a question → AI answers with your real data

Every AI query uses context injection — the system extracts column names, data types, summary statistics, sample rows, and null counts from your dataset, then sends that context alongside your question. The AI never guesses; it works with your actual numbers.

---

Tech Stack

Layer	Technology
Frontend	Streamlit
Data Processing	Pandas
AI (Production)	AWS Bedrock — Mistral 8B
AI (Local/Free)	Groq — LLaMA 3.3 70B
Storage	AWS S3 + local dual mode
Compute	AWS EC2
Auth	JWT + bcrypt
Database	SQLite + SQLAlchemy
Containers	Docker + docker-compose
CI/CD	GitHub Actions

---

Quick Start

Local Development (completely free)

git clone https://github.com/BurraRohan/Cloud-native-GenAI-CSV-Analytics-Platform.git
cd Cloud-native-GenAI-CSV-Analytics-Platform
 
python -m venv venv
source venv/bin/activate        # Mac/Linux
# venv\Scripts\activate          # Windows
 
pip install -r requirements.txt
 
cp .env.example .env
# Get a FREE Groq API key at https://console.groq.com (no credit card needed)
# Set GROQ_API_KEY in .env
 
streamlit run app.py
# Open http://localhost:8501

Local mode uses STORAGE_MODE=local and AI_MODE=groq. No AWS account needed — Groq's free tier gives 14,400 requests/day.

Docker

docker-compose up --build
# Access at http://localhost:8501

---

Features

Context-Injected AI Responses

The AI doesn't just receive your question — it receives your entire dataset's structure. build_context() extracts columns, types, df.describe() output, first 5 rows, and null counts. A few-shot system prompt teaches the model to respond with specific numbers, percentages, and actionable recommendations.

Role-Based Access Control

Three enforced roles with real feature restrictions — not cosmetic labels:

Feature	Admin	Analyst	Viewer
Dashboard	✅	✅	✅
Upload CSV	✅	✅	❌
Ask AI	✅	✅	✅
Load Datasets	✅	✅	✅
User Management	✅	❌	❌
Download PDF Report	✅	✅	✅

Dual Environment Modes

Switch between local and AWS with zero code changes — just toggle .env variables:

Mode	Storage	AI Provider	Cost
Local Dev	STORAGE_MODE=local	AI_MODE=groq	Free
AWS Production	STORAGE_MODE=s3	AI_MODE=bedrock	AWS pricing
Mixed	STORAGE_MODE=s3	AI_MODE=groq	S3 costs only

PDF Report Generation

Download AI insights as formatted PDF documents — useful for sharing analysis with non-technical stakeholders.

Security

• JWT authentication with bcrypt password hashing
• Role-based access control enforced server-side
• IAM least-privilege roles (EC2 → S3 + Bedrock only)
• API keys in environment variables, never hardcoded
• CSV input validation (type, size, readability checks)
• AES-256 server-side encryption on all S3 uploads
• HTTPS in production

Monitoring

AWS CloudWatch dashboard tracking S3 storage metrics, Bedrock invocation latency, and token usage across input/output counts.

Testing

87 tests across 5 modules covering authentication, file upload/validation, data processing, AI integration, and configuration. Run with pytest.

CI/CD

GitHub Actions pipeline triggered on every push to main — runs module import tests, verifies auth (hash + verify), and builds the Docker image.

---

AWS Deployment

1. Create S3 Bucket

Create a bucket named insightcloud-uploads in us-east-1 with public access blocked and AES-256 encryption enabled.

2. Create IAM Role

Create an EC2 role with a custom policy (deploy/iam_policy.json) granting least-privilege access to S3 and Bedrock only.

3. Launch EC2 and Deploy

# On a fresh Ubuntu 24.04 EC2 instance (t3.micro)
sudo apt update && sudo apt install -y docker.io docker-compose git
git clone https://github.com/BurraRohan/Cloud-native-GenAI-CSV-Analytics-Platform.git
cd Cloud-native-GenAI-CSV-Analytics-Platform
nano .env  # Configure your environment variables
sudo docker-compose up --build -d

Access at http://your-ec2-public-ip:8501

Security group should allow inbound on ports 22 (SSH) and 8501 (Streamlit).

---

Project Structure

├── app.py                 # Streamlit UI — login, dashboard, upload, AI query
├── auth.py                # JWT authentication — signup, login, token management
├── upload.py              # CSV upload & validation — local + S3 dual mode
├── processing.py          # Pandas data analysis & summary statistics
├── genai.py               # AI integration — Groq + Bedrock dual mode
├── database.py            # SQLAlchemy + SQLite setup
├── models.py              # User database model
├── config.py              # Centralized configuration & mode toggles
├── s3_utils.py            # AWS S3 helper functions
├── report.py              # PDF report generator
├── requirements.txt
├── pytest.ini
├── Dockerfile
├── docker-compose.yml
├── .env.example           # Environment variables template
├── tests/                 # 87 test cases across 5 modules
├── .streamlit/
│   └── config.toml        # Streamlit theme (warm gold #B88E23)
├── deploy/
│   ├── ec2_setup.sh       # EC2 bootstrap script
│   └── iam_policy.json    # Least-privilege IAM policy
└── .github/
    └── workflows/
        └── ci.yml         # GitHub Actions CI/CD pipeline

---

Future Improvements

• Migrate frontend to React for a smoother production experience
• Add support for Excel (.xlsx) uploads alongside CSV
• Implement GenAI response caching to reduce latency and cost
• Add auto-scaling with ECS/EKS for handling concurrent users
• Integrate SonarQube for automated code quality checks

---