Manufacturing Batch Clustering System

An end-to-end Machine Learning and web application for classifying industrial chemical production batches into operational performance clusters using K-Means clustering.

---

📌 Table of Contents

1. Overview
2. Project Architecture
3. Features
4. Machine Learning Model & Clusters
5. Project Structure
6. Setup & Installation
   • Prerequisites
   • Backend Setup (Flask API)
   • Frontend Setup (React + Vite)
7. API Endpoints

---

🔬 Overview

This project classifies chemical batch manufacturing data in real-time. By analyzing 10 key production variables (such as temperature, pressure, reaction time, yield, purity, and energy consumption), an unsupervised K-Means Clustering model assigns each batch to one of three operational quality categories.

---

🏗 Project Architecture

The application is structured into a modern two-tier architecture:

• Frontend (React + Vite + CSS3): A clean, responsive industrial UI that collects process input variables and presents real-time cluster classifications.
• Backend (Python Flask + Scikit-Learn): REST API service that accepts JSON payload, scales inputs with a pre-trained StandardScaler, and predicts the cluster using a trained KMeans model (joblib).

graph TD
    UI[React Frontend UI] -->|POST /get_clusters| API[Flask Backend API]
    API --> SCALER[StandardScaler - scaler.pkl]
    SCALER --> MODEL[K-Means Model - kmeans_model.pkl]
    MODEL -->|Cluster ID| API
    API -->|JSON Response: Cluster, Name, Description| UI

Loading

---

✨ Features

• Real-time Batch Classification: Predicts cluster membership instantly upon form submission.
• Multi-variable Parameter Input: Evaluates 10 distinct physical & chemical metrics:
  • Temperature (°C) & Pressure (atm)
  • Reaction Time (min) & Cooling Time (min)
  • Catalyst Loading (%) & pH Level
  • Yield (%) & Purity (%)
  • Energy (kWh) & Mixing Speed (RPM)
• Pre-trained ML Models: Uses Scikit-Learn KMeans and StandardScaler loaded via joblib.
• Clean Industrial UI: Responsive React interface designed with an emphasis on clarity and accessibility.

---

📊 Machine Learning Model & Clusters

The model segments production batches into three distinct operational profiles:

Cluster ID	Cluster Name	Profile & Description
0	Optimised Production	High purity and yield accompanied by low operating costs.
1	Standard Production	Moderate purity and yield with baseline production costs.
2	High Risk Production	Low purity and yield paired with high energy/operating costs.

---

📁 Project Structure

manifacturing_batch_clustering_system/
│
├── backend/                  # Flask REST API
│   ├── app.py                # Main Flask application & routes
│   ├── requirments.txt       # Python dependencies
│   └── model/                # Pre-trained ML models
│       ├── kmeans_model.pkl  # Trained KMeans model
│       └── scaler.pkl        # Fitted StandardScaler object
│
├── batch-clustering/         # React Frontend (Vite)
│   ├── src/
│   │   ├── App.jsx           # Parent App component
│   │   ├── FormData.jsx      # Batch parameter input form & result display
│   │   ├── index.css         # Styling rules & design system
│   │   └── main.jsx          # Entry point
│   ├── package.json          # Node dependencies
│   └── vite.config.js        # Vite configuration
│
├── data/                     # Raw & processed manufacturing datasets
└── clustering_system.ipynb   # Jupyter Notebook for EDA, preprocessing, and model training

---

🚀 Setup & Installation

Prerequisites

• Python 3.8+
• Node.js 16+ & npm

---

1. Backend Setup (Flask API)

1. Navigate to the backend directory:

   cd backend

2. Create and activate a Python virtual environment:

   python -m venv venv
   # On Windows:
   venv\Scripts\activate
   # On macOS/Linux:
   source venv/bin/activate

3. Install required dependencies:

   pip install flask flask-cors joblib pandas scikit-learn

4. Start the Flask API server:

   python app.py

   The server runs locally at http://127.0.0.1:5001/

---

2. Frontend Setup (React + Vite)

1. Open a new terminal and navigate to the frontend directory:

   cd batch-clustering

2. Install Node packages:

   npm install

3. Start the Vite development server:

   npm run dev

4. Open your browser and navigate to http://localhost:5173/ (or the URL outputted by Vite).

---

🔌 API Endpoints

GET /

• Description: Health check endpoint.
• Response: "Manufacturing API is running Successfully"

POST /get_clusters

• Description: Accepts batch metrics and returns predicted cluster info.
• Request Body Example:

  {
    "temperature": 129,
    "pressure": 2.5,
    "reaction_time": 45,
    "catalyst_loading": 1.2,
    "pH": 7.0,
    "yield_pct": 94.5,
    "purity_pct": 98.2,
    "energy": 320,
    "cooling_time": 15,
    "mixing_speed": 1500
  }

• Response Example:

  {
    "status": "success",
    "cluster": 0,
    "name": "Optimised Production",
    "description": "High Purity and Yield with low Operating Cost"
  }

---

📄 License

This repository is maintained for industrial manufacturing process analytics.