AutoML - An Intelligent ML Analytics Engine

This system ingests a CSV dataset, performs automated preprocessing, exploratory data analysis, model training, evaluation, and generates a polished analytical report. It produces performance metrics and visual insights — including EDA plots and model evaluation charts — through a simple web interface. Designed for students, analysts, and data teams who want fast, structured insights without manual setup.

---

Key Features

Automated Analysis Suite

• Data Preprocessing : Missing-value handling, Numeric + categorical feature processing, Date column detection & conversion, Automatic dataset cleaning, Outputs a clean processed.csv

• Exploratory Data Analysis (EDA): Summary statistics (describe), Missing-value profiling, Histogram visualizations for numeric features, Correlation heatmap, Safe filename handling, All plots saved under runs/<run_id>/eda_plots/

• AutoML Model Training : Automatically detects problem type (regression/classification), Trains multiple baseline models, Selects best model automatically, Saves model as best_model.pkl

• Model Evaluation & Reporting : Regression Metrics - R² Score, RMSE, MAE, MSE | Classification Metrics - Accuracy, Confusion Matrix | Evaluation Plots - Actual vs Predicted, Residual Plot, Error Distribution | PDF final report generation | Results stored in structured folders per run

---

User-Friendly Interface

• Web UI built with Flask + vanilla HTML/JS
• Upload CSV & select target column
• Click once to run the full pipeline
• Live status updates at each step
• View metrics, charts & download report
• Runs stored safely by unique run_id

---

Technology Stack

• Backend : Python, Flask (REST API), scikit-learn (ML models & metrics), pandas / numpy (data processing), seaborn / matplotlib (visualization), joblib (model persistence)

• Frontend : HTML, CSS, JavaScript (Fetch API)

• Architecture & Infrastructure : Thread-based async training, Structured run directories, CORS enabled frontend-backend communication, Headless plotting via matplotlib.use("Agg")

---

Website Overview

---

Quick Start

Clone the repository and install dependencies:

git clone https://github.com/BhaveshBhakta/Intelligent-ML-Analytics-Engine.git
cd Intelligent-ML-Analytics-Engine
pip install -r requirements.txt

Run the backend:

python -m backend.app

Open the UI:

http://localhost:5000

All outputs are saved automatically under:

runs/<RUN_ID>/

---

High-Level Architecture

User (Browser, CSV Upload)
        ↓
     Flask API
        ↓
 ┌────────────── Pipeline ───────────────┐
 │ Upload & Run Creation                 │
 │ Data Preprocessing                    │
 │ Exploratory Data Analysis (EDA)       │
 │ AutoML Model Training                 │
 │ Model Evaluation + Charts             │
 │ PDF Report Generation                 │
 └───────────────────────────────────────┘
        ↓
  UI Dashboard + Exportable Results

---

Roadmap & Future Enhancements

• Automated hyperparameter tuning
• Explainability using SHAP / LIME
• Outlier detection & handling
• Time-series forecasting support
• Model comparison dashboard
• Authentication & multi-user runs
• Cloud deployment template