ML Algorithm Visualizer with Groq AI

A comprehensive machine learning visualization tool that lets you explore datasets, train multiple algorithms, and get AI-powered insights through Groq integration.

What it does

This app makes machine learning accessible by providing an intuitive interface to:

• Upload and explore your datasets with interactive visualizations
• Train and compare 9+ different ML algorithms (both classification and regression)
• Get AI-powered explanations and recommendations for your models
• Visualize model performance with charts, confusion matrices, and learning curves

Architecture

┌─────────────────────────────────────────────────────────────────┐
│                        Streamlit Frontend                       │
│                         (app.py)                               │
└─────────────────────┬───────────────────────────────────────────┘
                      │
┌─────────────────────▼───────────────────────────────────────────┐
│                   UI Components                                 │
│                 (ui_components.py)                             │
│  ┌─────────────┐ ┌─────────────┐ ┌─────────────┐ ┌───────────┐ │
│  │    Data     │ │     ML      │ │   Model     │ │    AI     │ │
│  │Visualization│ │ Algorithms  │ │ Comparison  │ │ Assistant │ │
│  │     Tab     │ │     Tab     │ │     Tab     │ │    Tab    │ │
│  └─────────────┘ └─────────────┘ └─────────────┘ └───────────┘ │
└─────┬───────────────┬───────────────┬───────────────┬───────────┘
      │               │               │               │
┌─────▼─────┐ ┌───────▼──────┐ ┌──────▼──────┐ ┌─────▼─────────┐
│   Data    │ │      ML      │ │   Session   │ │     Groq      │
│  Handler  │ │  Algorithms  │ │    State    │ │   Assistant   │
│           │ │              │ │   Manager   │ │               │
│• Dataset  │ │• 9 Regression│ │• Model      │ │• Chat         │
│  Loading  │ │  Algorithms  │ │  Results    │ │• Analysis     │
│• Preprocessing│ │• 6 Classification│ │• Dataset    │ │• Recommendations│
│• Visualization│ │  Algorithms  │ │  Storage    │ │• Explanations │
│• Statistics│ │• Training    │ │• UI State   │ │               │
│           │ │• Evaluation  │ │             │ │               │
└───────────┘ └──────────────┘ └─────────────┘ └───────────────┘
      │               │                               │
┌─────▼─────┐ ┌───────▼──────┐                 ┌─────▼─────────┐
│  Plotly   │ │   Scikit-    │                 │   Groq API    │
│  Charts   │ │    Learn     │                 │   (External)  │
│           │ │              │                 │               │
│• Correlation│ │• Models     │                 │• gemma2-9b-it │
│• Distribution│ │• Metrics    │                 │• Chat         │
│• Pairplots │ │• Validation │                 │• Analysis     │
│• Heatmaps  │ │• Visualization│                │               │
└───────────┘ └──────────────┘                 └───────────────┘

Core Components

Data Handler (data_handler.py)

Manages all data operations including loading sample datasets (Iris, Wine, Breast Cancer), preprocessing features, handling categorical variables, and creating visualizations like correlation heatmaps and distribution plots.

ML Algorithms (ml_algorithms.py)

Contains the MLAlgorithmVisualizer class that handles:

• Regression: Linear, Ridge, Lasso, Elastic Net, KNN, SVR, Decision Tree, Random Forest, Gradient Boosting
• Classification: Logistic Regression, KNN, SVM, Decision Tree, Random Forest, Gradient Boosting
• Model training, evaluation, and visualization (learning curves, feature importance, decision boundaries)

Groq AI Assistant (groq_assistant.py)

Integrates with Groq's API to provide:

• Dataset analysis and insights
• Algorithm recommendations based on your data
• Model results explanations in plain English
• Interactive chat for ML questions

UI Components (ui_components.py)

Four main tabs that provide the user interface:

1. Data & Visualization: Upload datasets, explore data with interactive charts
2. ML Algorithms: Configure and train models with real-time results
3. Model Comparison: Compare performance across different algorithms
4. AI Assistant: Get AI-powered insights and recommendations

Getting Started

1. Install dependencies:

pip install streamlit pandas plotly scikit-learn groq python-dotenv

2. Set up Groq API (optional but recommended): Create a .env file in the project root:

GROQ_API_KEY=your_groq_api_key_here

Or add it to Streamlit secrets in .streamlit/secrets.toml:

GROQ_API_KEY = "your_groq_api_key_here"

3. Run the app:

streamlit run app.py

Features

Data Exploration

• Upload CSV files or use built-in sample datasets
• Interactive correlation heatmaps and distribution plots
• Dataset statistics and missing value analysis
• Pairplot visualizations with target coloring

Machine Learning

• Automatic problem type detection (classification vs regression)
• 15 different algorithms to choose from
• Cross-validation scoring and performance metrics
• Visual model evaluation (confusion matrices, prediction plots)
• Feature importance analysis for tree-based models
• Learning curves to understand model behavior

AI Integration

• Chat with AI about your data and models
• Get personalized algorithm recommendations
• Understand model results with AI explanations
• Compare multiple models with AI insights

Sample Datasets

The app includes three classic ML datasets:

• Iris: Flower classification (4 features, 3 classes)
• Wine: Wine quality classification (13 features, 3 classes)
• Breast Cancer: Cancer diagnosis (30 features, 2 classes)

Technical Details

Built with:

• Streamlit for the web interface
• Scikit-learn for machine learning algorithms
• Plotly for interactive visualizations
• Groq API for AI-powered insights
• Pandas for data manipulation

The app uses session state to maintain data and model results across tabs, allowing you to train multiple models and compare them seamlessly.

Tips

• Start with the sample datasets to get familiar with the interface
• Try different algorithms on the same dataset to see which performs best
• Use the AI assistant to understand why certain algorithms work better for your data
• The model comparison tab helps you make informed decisions about which algorithm to use

Perfect for data scientists, students, and anyone wanting to understand machine learning through hands-on experimentation.