📊 DataLens AI

Intelligent Data Analytics Agent

An autonomous AI-powered data analytics system that transforms raw datasets into professional visualizations and interactive dashboards

🚀 Live Demo (V1) • 📺 Video Demo •

Capstone Project for Google's 5-Day AI Agents Intensive Course

📖 Table of Contents

• 🎯 Problem & Solution
• 🏗️ System Architecture
• ✨ Core Capabilities
• 🚀 Quick Start
• 🎨 Features
• 🔧 Technical Stack
• 📊 Output Deliverables
• 📈 Use Cases
• 🛡️ Security
• 📚 Project Structure

🎯 Problem & Solution

The Challenge

Modern data analysis faces critical barriers:

• Complexity: Multiple tools required for cleaning, analysis, and visualization
• Technical Skills: Demands expertise in Python, pandas, and visualization libraries
• Time Investment: Manual processes consume hours of productive time
• Accessibility: Non-technical users locked out of advanced analytics
• Inconsistency: Variable quality based on individual expertise

Our Solution

DataLens AI democratizes data analysis through AI automation:

Raw Data → AI Processing → Professional Insights
   ↓            ↓                    ↓
Upload → Gemini Analysis → Interactive Dashboard

Key Benefits:

• 🤖 AI-Driven: Leverages Google's Gemini API for intelligent processing
• ⚡ Fast: Hours of work reduced to minutes
• 🎯 Complete: End-to-end pipeline in a single notebook
• 🚀 No-Code: Upload and process without manual coding
• 📊 Professional: Publication-quality visualizations

🏗️ System Architecture

Pipeline Workflow

graph LR
    A[🔧 Setup Environment] --> B[🤖 Initialize Gemini AI]
    B --> C[📁 Load Data]
    C --> D[🧠 AI Analysis & Cleaning]
    D --> E[📊 Generate Visualizations]
    E --> F[📈 Interactive Dashboard]
    
    style A fill:#e3f2fd,stroke:#1976d2,stroke-width:3px,color:#000000
    style B fill:#f3e5f5,stroke:#7b1fa2,stroke-width:3px,color:#000000
    style C fill:#fff3e0,stroke:#f57c00,stroke-width:3px,color:#000000
    style D fill:#e8f5e9,stroke:#388e3c,stroke-width:3px,color:#000000
    style E fill:#fce4ec,stroke:#c2185b,stroke-width:3px,color:#000000
    style F fill:#e0f2f1,stroke:#00796b,stroke-width:3px,color:#000000

Loading

Component Architecture

┌─────────────────┐    ┌──────────────────┐    ┌──────────────────┐
│   Data Input    │    │  AI Processing   │    │ Output Generation│
│                 │    │                  │    │                  │
│ • CSV/Excel     │───▶│ • Gemini AI     │───▶│ • Visualizations │
│ • Raw Data      │    │ • Analysis       │    │ • Dashboard      │
│ • File Upload   │    │ • Code Generation│    │ • Reports        │
└─────────────────┘    └──────────────────┘    └──────────────────┘
         │                        │                        │
         └────────────────────────┼────────────────────────┘
                                  │
                         ┌──────────────────┐
                         │ Data Processing  │
                         │                  │
                         │ • Cleaning       │
                         │ • Transformation │
                         │ • Encoding       │
                         └──────────────────┘

✨ Core Capabilities

🤖 AI-Driven Intelligence Gemini API for automated quality assessment and insights generation	🧹 Smart Data Cleaning Intelligent cleaning code generation based on data profiling	📊 Advanced Visualizations 10+ chart types with professional styling and interactivity
📈 Interactive Dashboard Real-time filters, KPI cards, and auto-updating charts	🏭 Production-Ready ML-ready datasets with encoding and standardization	📁 Seamless Upload Interactive widget supporting CSV and Excel formats

🚀 Quick Start

Prerequisites

Requirement	Version	Status
Python	3.8+	✅ Required
Google Colab	-	🌟 Recommended
Gemini API Key	-	🔑 Required

Installation

pip install pandas numpy matplotlib seaborn plotly scikit-learn ipywidgets \
            jsonschema google-generativeai google-auth google-auth-oauthlib \
            openpyxl xlrd jupyterlab

Setup Steps

1. Access the Notebook

# Open in Google Colab
File → Upload → Select the .ipynb notebook

2. Configure API Key

# Add Gemini API key to Colab Secrets
# 1. Click 🔑 in left sidebar
# 2. Add new secret: GEMINI_API_KEY = "your_api_key"

3. Run the Pipeline

# Execute cells sequentially:
# Cells 1-2:  Environment setup
# Cells 3-4:  AI initialization  
# Cell 5:     Data upload
# Cells 6-9:  AI cleaning
# Cells 10-14: Visualizations
# Cells 15-17: Dashboard
# Cells 18-19: Reports

Usage Example

🔧 Initialize Environment

# Cell 1: Install dependencies
!pip install pandas numpy matplotlib seaborn plotly scikit-learn ipywidgets \
            jsonschema google-generativeai --quiet

⏱️ ~2 minutes

🤖 Initialize Gemini AI

# Cell 3-4: Configure API
from google.colab import userdata
import google.genai as genai

api_key = userdata.get("GEMINI_API_KEY")
client = genai.Client(api_key=api_key)

⏱️ ~30 seconds

📁 Load Dataset

# Cell 5: Upload and analyze
df = upload_dataset()
dataset_summary = generate_dataset_summary(df)

⏱️ Variable (depends on file size)

🧹 AI-Powered Cleaning

# Cell 6-7: Automated cleaning
cleaning_prompt = build_cleaning_prompt(dataset_summary)
cleaning_output = ask_gemini_cleaning(cleaning_prompt)

⏱️ ~1 minute

📊 Generate Visualizations

# Cell 10-14: Create charts
viz_code = prompt_gemini(viz_prompt)
exec(viz_code)

# Cell 15-17: Build dashboard
dashboard_code = prompt_gemini(dash_prompt)
exec(dashboard_code)

⏱️ ~2 minutes

🎨 Features

🔍 Automated Data Analysis

• Comprehensive Summary: Statistical metrics, missing values, data type profiling
• AI Quality Assessment: Gemini-powered evaluation
• Column-wise Analysis: Detailed numeric and categorical insights

🧹 Smart Data Cleaning

Feature	Description	Status
Missing Value Detection	Automatic identification and handling	✅
Outlier Management	99th percentile statistical capping	✅
Data Normalization	Column standardization and value scaling	✅
Categorical Encoding	One-hot encoding for ML readiness	✅
Negative Value Handling	Automatic conversion to absolute values	✅

📊 Visualization Suite

📊 Chart Types • Histograms • Bar charts • Line charts • Scatter plots • Box plots • Heatmaps • Pie charts • Correlation matrices • Violin plots • Area charts

🎛️ Interactive Features • Real-time filtering • KPI cards • Multi-select widgets • Auto-updating charts • Dynamic interactions • Responsive design

✨ Professional Quality • Custom styling • Proper titles • Axis labels • Legends • Color schemes • Export-ready

🤖 AI Integration

🧠 Gemini 2.5 Pro Advanced data analysis and cleaning recommendations 🔴 Thorough & Comprehensive

⚡ Gemini 2.5 Flash Fast visualization code generation 🟢 Quick & Efficient

🔧 Technical Stack

Core Dependencies

📊 Data Processing pandas # Data manipulation numpy # Numerical operations 📈 Visualization matplotlib # Static plots seaborn # Statistical graphics plotly # Interactive charts 🤖 Machine Learning scikit-learn # Preprocessing & encoding

🧠 AI Integration google-generativeai # Gemini API 🎛️ Interactive Components ipywidgets # Dashboard widgets ✅ Validation jsonschema # Data validation

Notebook Structure

📓 Cells 1-2:   Environment Setup (Dependencies)
🤖 Cells 3-4:   AI Initialization (Gemini Config)
📁 Cell 5:      Data Loading (Upload & Profile)
🧹 Cells 6-9:   AI Cleaning (Quality Improvement)
📊 Cells 10-14: Visualization (Chart Generation)
📈 Cells 15-17: Dashboard (Interactive Interface)
📋 Cells 18-19: Reporting (Insights & Recommendations)

📊 Output Deliverables

1️⃣ Cleaned Dataset ML-ready with encoding and standardization

2️⃣ Visualizations 10+ professional charts

3️⃣ Interactive Dashboard Real-time analytics with KPIs

4️⃣ Analysis Report Automated insights & recommendations

📈 Use Cases

💼 Business Intelligence Sales analysis & forecasting Performance tracking KPI monitoring & dashboards Revenue analysis Market trend identification 🔬 Data Science Automated ETL pipelines Feature engineering Model preparation Data preprocessing Exploratory data analysis

📊 Research Analytics Statistical analysis Correlation studies Pattern recognition Hypothesis testing Trend analysis 📋 Reporting Automation Automated report generation Executive dashboards Periodic reporting Stakeholder presentations Business intelligence insights

🛡️ Security

╔═══════════════════════════════════════════════════════════╗
║               SECURITY & PRIVACY MEASURES                 ║
╠═══════════════════════════════════════════════════════════╣
║  ✅  Secure API Handling                                  ║
║      → API keys stored in Colab secrets                   ║
║                                                           ║
║  ✅  No Hardcoded Credentials                             ║
║      → Secure authentication practices                    ║
║                                                           ║
║  ✅  Data Privacy                                         ║
║      → Local processing without external transmission     ║
╚═══════════════════════════════════════════════════════════╝

📚 Project Structure

DataLens-AI-Intelligent-Data-Analytics-Agent/
│
└── 📊 DataLens AI - Intelligent Data Analytics Agent.ipynb
    (Version 2 - Optimized for Google Colab)

🚀 Deployment

🌐 Version 1 - Live Demo Deployed on Hugging Face Spaces

📦 Version 2 - Current Available in this Repository

🚨 Important Notes

⚠️  REQUIREMENTS
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

✓ Google Colab environment recommended
✓ Gemini API key configured in Colab secrets
✓ Supports CSV and Excel file formats
✓ Automatic dependency installation

🎓 Capstone Project

Google's 5-Day AI Agents Intensive Course

Built Using

Transform your data into insights with AI ✨

Made by Adinath Somnath Jagtap & Prajwal Ashok Zolage