Media Reaction Simulator 🚀📊🎥

Expanded Overview 🌐✨

This repository now encompasses a specialized implementation for testing propositions from "The Oxford Handbook of Entertainment Theory" by Vorderer and Klimmt. Our pipeline, incorporating advanced scraping, metadata extraction, and feature analysis, integrates with a Large Language Model (LLM) to simulate user reactions to YouTube Shorts. This approach facilitates an in-depth exploration of media entertainment theories, particularly focusing on user control and affective responses.

Features and Implementation 🌟🔧

• Topic-Based Scraping: 🌐 Refined to target videos aligning with theoretical concepts under study.
• In-Depth Metadata Extraction: 📝 Now includes viewer engagement metrics and additional relevant parameters.
• Feature Extraction: 🔍
  • Words Per Minute (WPM) Analysis: Analyzing speech variation patterns over time.
  • Emotion Extraction: Utilizing NLP techniques for nuanced emotional expression analysis.
• LLM-Powered Simulation and Analysis: 🧠 Replicating and analyzing studies, focusing on:
  • Proposition 2.1 Analysis: Examining the relationship between control capacity, motivation, and affective reactions.
  • Proposition 2.2 Exploration: Investigating how control factors influence reappraisal of affective responses and goal prioritization.

Case Study: Testing Propositions 2.1 & 2.2 📚🔬

We conducted a case study focusing on Propositions 2.1 and 2.2 from Media Entertainment as Guilty Pleasure? The Appraisal of Media Use, Self-Control, and Entertainment (AMUSE) Model by Reinecke and Meier

• Proposition 2.1: "Control capacity and control motivation negatively predict the intensity of both positive and negative primary affective reactions to within-media events."
• Proposition 2.2: "Control capacity and motivation moderate the reappraisal of primary affective responses to within-media events, with low self-control favoring short-term mood management goals, and high self-control favoring higher-order goals."

Methodology:

1. Data Collection: Gathered a diverse range of YouTube Shorts.
2. Feature Analysis: Applied emotion and speech pattern analysis.
3. LLM Simulation: Simulated user reactions based on control capacity and motivation levels.
4. Comparative Analysis: Compared LLM-generated reactions with theoretical expectations.