Twitch Chat Real-Time Analyzer

A real-time analytics application designed to capture and process Twitch IRC chat messages, identifying viewer engagement patterns and providing insights to optimize streamer interactions.

Features

• Real-Time Chat Capture: Connects to Twitch IRC to capture chat messages as they happen.
• Asynchronous Processing: Utilizes an asynchronous pipeline for efficient handling of incoming messages.
• NLP Analysis:
  • Sentiment Analysis: Analyzes the sentiment (positive, negative, neutral) of chat messages using NLTK.
  • Keyword Extraction: Identifies frequently used keywords and topics in the chat.
• Emote Analysis:
  • Detects standard Twitch emotes, FFZ, and 7TV emotes within messages.
  • Analyzes emote usage frequency and sentiment (based on a custom dataset).
• Real-Time Dashboard: (Planned) A web-based dashboard (using React and WebSockets) to visualize:
  • Overall chat sentiment trends over time.
  • Top keywords and their frequency.
  • Most used emotes and their sentiment scores.
  • Viewer engagement metrics.
• Alerts: (Planned) Configurable alerts for significant shifts in chat sentiment, specific keyword usage spikes, or unusual activity patterns.
• Streamer Input UI: A simple interface (in the frontend) for the streamer to input the target Twitch channel and start/stop the analysis.

Project Structure

• backend/: Contains the Python FastAPI application responsible for:
  • Connecting to Twitch IRC (twitch_irc.py).
  • Processing messages (NLP: nlp_processor.py, Emotes: emote_handler.py).
  • Managing WebSocket connections for real-time frontend updates (websocket_manager.py).
  • The main application logic (main.py).
• frontend/: Contains the React application for the user interface and dashboard.
  • Displays real-time analytics received via WebSockets.
  • Provides controls for starting/stopping analysis.
• emoji_sentiment_scores.csv: A custom dataset mapping 7TV emotes to sentiment scores. (See Note below)
• .gitignore: Specifies intentionally untracked files that Git should ignore.
• requirements.txt: Lists the Python dependencies for the backend.
• README.md: This file.

Setup

Prerequisites:

• Python 3.8+
• Node.js and npm (or yarn)

Backend:

# Navigate to the backend directory
cd backend

# Create and activate a virtual environment (recommended)
python -m venv venv
# Windows Powershell:
.\venv\Scripts\Activate.ps1
# Bash/Zsh:
# source venv/bin/activate

# Install Python dependencies
pip install -r requirements.txt

# Download necessary NLTK data models
python -m nltk.downloader -d ./nltk_data punkt averaged_perceptron_tagger stopwords wordnet omw-1.4

# Create a .env file based on .env.example (if provided)
# and add your Twitch credentials (Username, OAuth Token)
# Example .env:
# TWITCH_USERNAME="your_bot_username"
# TWITCH_OAUTH_TOKEN="oauth:your_oauth_token"
# TARGET_CHANNEL="target_streamer_channel"
# CLIENT_ID="your_client_id"        # Optional: Needed for some API calls if implemented
# CLIENT_SECRET="your_client_secret" # Optional: Needed for some API calls if implemented
# OPENAI_API_KEY="your_openai_key"    # If using OpenAI for advanced NLP

# Run the backend server
uvicorn main:app --reload --host 0.0.0.0 --port 8000

Frontend:

# Navigate to the frontend directory
cd frontend

# Install Node.js dependencies
npm install
# or: yarn install

# Run the frontend development server
npm run dev
# or: yarn dev

# The frontend will typically be available at http://localhost:5173

Technology Stack

• Backend: Python, FastAPI, Uvicorn, websockets, NLTK, python-dotenv
• Frontend: React, TypeScript, Vite, Socket.IO Client (or native WebSocket API)
• Real-time Communication: WebSockets
• NLP: NLTK (Natural Language Toolkit)

Note on emoji_sentiment_scores.csv

This project utilizes a custom dataset (emoji_sentiment_scores.csv) to assign sentiment scores to specific 7TV emotes.