Complaint Category Classifier — AI/ML

An NLP-based machine learning model that automatically classifies customer complaints into predefined categories, enabling faster and more accurate complaint routing.

📋 Overview

Manual complaint sorting is slow and inconsistent. This project uses natural language processing to automatically categorize incoming complaints — helping route them to the right department faster and reducing manual workload.

✨ Features

• Classifies complaints into 4 categories: Water, Electricity, Roads, Sanitation
• TF-IDF vectorization + Naive Bayes classification model
• FastAPI backend for serving predictions
• Streamlit dashboard for an interactive user interface
• Admin correction panel to fix misclassified complaints
• Prediction history export for tracking and analysis

🛠️ Tech Stack

• Language: Python
• ML/NLP: scikit-learn (TF-IDF, Naive Bayes)
• Backend: FastAPI
• Frontend: Streamlit
• Data handling: Pandas

🚀 How to Run

1. Clone the repository

   git clone https://github.com/maha-434/complaint-classifier.git
   cd complaint-classifier

2. Install dependencies

   pip install -r requirements.txt

3. Train the model

   python train.py

4. Run the FastAPI backend

   uvicorn app:app --reload

5. Run the Streamlit dashboard (in a separate terminal)

   streamlit run dashboard.py

📂 Project Structure

complaint-classifier/
├── data.csv           # Labeled complaint dataset
├── train.py            # Model training script
├── app.py               # FastAPI backend
├── dashboard.py    # Streamlit frontend
└── requirements.txt

📈 Future Improvements

• Expand dataset for improved model accuracy
• Add train/test split with evaluation metrics
• Support for additional complaint categories
• Multi-language complaint support

👤 Author

Mahalakshmi M
LinkedIn · GitHub