Comment Sentiment Detector

A web application powered by Machine Learning capable of determining the sentiment of a comment.

📚 Table of contents

• Technical stack
  • Frontend
  • Backend
• Features
• Run with Docker
• Installation
• Usage
• Todo
• License

🛠 Technical stack

🖍️ Frontend

• Framework: Bootstrap

⚙ Backend

• Programming language(s): Python
• App framework: Flask
• Distributed memory caching: Memcached
• Data processing and Machine Learning: sklearn, numpy, pandas
• HTTP Load Balancing: NGINX
• Deployment: Docker, AWS

🚀 Features

• Preprocess a data file of natural language sentences
• Train a Machine Learning model based on the processed data
• Evaluate natural language sentences to extract sentiments and confidence scores

🐳 Run with Docker

$ docker-compose up

The application can be found at http://localhost:5000/

⬇ Installation

0. Prerequisites

Make sure you have installed these following prerequisites on your computer:

• Python 3.6 or later. You can download different versions of Python here: http://www.python.org/getit/
• pip
• virtualenv

1. Install and activate virtual environment

$ virtualenv venv --python=python3.8
$ source venv/bin/activate

2. Install requirements

pip install -r requirements.txt

3. Run the application:

$ python run.py

🤟 Usage

Comment Sentiment Detector can be used either as a web application, APIs, or scripts.

• Web application: The application can be found at http://localhost:5000/
• APIs: APIs for backend usage can be found here. For example, send a POST request to /api/v1/sentiments/predict to evaluate sentiment of a text, example request body:

{
  "text": "this is awesome!"
}

A JSON object with the detected sentiment and confidence score will be returned.

• Scripts: preprocess data, train model

📋 To-do

• Implement APIs for preprocessing data and training model
• Provide detailed API specs for backend usage

📄 License

MIT