Analyzing user comments on social media platforms is essential for understanding user responses and preferences, as well as for enhancing future media production. In this preliminary research, we proposed an ensemble sentiment scoring system that integrates three subsystems: a dictionary-based system, a large language model (LLM)-based system, and a local classification system. By aggregating votes from each subsystem and applying dynamic weighting criteria, a final sentiment score is reached. We conducted a small experiment to evaluate the effectiveness and accuracy of the system. Based on our current design and findings, we discussed several potential directions for future research.