Hate Speech Detection

Introduction

This project focuses on building a deep learning model to detect hate speech in twitter data. Utilizing a robust CI/CD pipeline with CircleCI and Google Cloud Platform (GCP), the project aims to deliver an efficient and scalable solution for hate speech detection.

Problem Statement

Hate speech on social media and other platforms is a growing concern, posing significant challenges to maintaining a safe online environment. This project aims to develop an automated system to accurately identify and filter out hate speech from text, thereby promoting a healthier and more respectful digital space.

Dataset

Dataset using Twitter data, isused to research hate-speech detection. The text is classified as: hate-speech, offensive language, and neither. Due to the nature of the study, it’s important to note that this dataset contains text that can be considered racist, sexist, homophobic, or generally offensive.

Link for dataset

Prerequisites

• Python 3.9 or higher
• Fast API
• Docker
• Circle CI
• Google Cloud SDK
• Google Cloud Platform(GCP) Account

Project CI/CD Workflow

Installation

1. Clone the repository

   git clone https://github.com/Komalsai234/Hate-Speech-Detection.git

2. Create a new conda environment

   conda create -p hate-speech python==3.9 -y

3. Install dependencies

   pip install -r requirements.txt

4. Configure the Google CLI in the terminal

Install the gCloud CLI and then run below command in terminal then enter the required account information

  gcloud init

5. Execute the Project pipeline

   Run the main.py script to execute the Pipeline

   python main.py

6. Start the Fast API Application

   Run the app.py script to start the Streamlit web application.

   python app.py

7. Interact with the Application

   Open your web browser and go to the local Fast API URL (usually http://localhost:8080)

Guide for Integrating CircleCI with GCP Virtual Machine