How it works
- The Web App starts the CommentsProvider, WebServer and the Sentiment Analysis Consumer each on a different Thread.
- The CommentsProvider starts the YoutubeScraper that fetches videos using a Search Term then monitors the videos.
- The Spark App loads the trained pickled models, starts a Kafka Consumer that listens to incoming comments, performs sentiment analysis then sends the results using a Kafka Producer (The WebApp will then send the results to clients connected in the WebServer).
- The HTML app connects to the WebServer, listens to incoming results and shows them in the page (it will also fetch the video's title if needed).
- Start the Kafka Zookeeper and Server.
- Start the web_app.py.
- Open the index.html in the browser.
- Start the sentiment_analysis.py (using SparkSubmit).
- Wait for the analysis to show in the page..