📊 ACAML: Adaptive Constraint-Aware AutoML Web Application

Welcome to ACAML, an innovative web-based AutoML application built with Streamlit. This project demonstrates my ability to design, implement, and deploy real-world machine learning solutions with user-friendly interfaces and powerful automation features.

---

🚀 Project Overview

ACAML empowers users to:

• Upload their own datasets in CSV format
• Automatically detect the appropriate machine learning task (classification or regression)
• Train the best model using FLAML with time-budgeted optimization
• Display key performance metrics (Accuracy or R² Score) clearly and intuitively
• Visualize model interpretability through SHAP feature importance plots

👉 Live Demo of the App

This app is designed to democratize machine learning for users of all skill levels, making it easy to explore, train, and interpret models interactively.

---

🛠️ Key Features

• Dynamic Task Detection: Automatically classifies datasets as regression or classification based on target variable type and cardinality.
• Automated Model Selection: Leverages FLAML to choose the best model and optimize hyperparameters within a specified time budget.
• User-Centric Interface: Clean, tabbed layout with separate sections for configuration, results, and explainability.
• Model Interpretability: Integrated SHAP visualizations to understand feature importance and build trust in the model.
• Flexible Deployment: Fully containerized and deployable on Streamlit Cloud or locally.

---

🔍 Technologies Used

• Python
• Streamlit
• FLAML
• SHAP
• scikit-learn
• pandas, numpy, matplotlib

---

📸 Screenshots

Screenshots of key app pages (Upload, Results, Explainability) are included in the screenshots/ folder. They showcase the intuitive user interface and visualizations.

---

📈 Why This Matters

As a Data Analyst/Machine Learning Engineer, I built ACAML to showcase my skills in:

• Designing user-friendly data science tools
• Implementing end-to-end ML workflows from preprocessing to model interpretability
• Deploying and maintaining ML applications in real-world scenarios

This project demonstrates my ability to bridge the gap between technical implementation and user experience, ensuring that complex machine learning processes are accessible and understandable to non-technical users.

---

📝 Getting Started

Clone the repo and run locally:

git clone https://github.com/Chetnas8/acaml-web-app.git
cd acaml-web-app
pip install -r requirements.txt
streamlit run app.py