FISAToon

Project Duration: Mar 11 – Mar 13, 2025

A platform for sentiment analysis and popularity prediction based on webtoon comment data

• Tech Stack: FastAPI, MongoDB, Selenium, HTML/CSS/JS
• Team & Role: 3 members (My Role – Fullstack & Crawling, Teammate 1 – Sentiment Analysis Model, Teammate 2 – Popularity Prediction & Visualization)
• GitHub: FISAToon

📌 Project Overview
• A data analysis project that performs sentiment analysis on Naver Webtoon comments and predicts popularity trends based on the results.

🎯 Motivation & Goals
• Explore the impact of emotional changes in comments on webtoon ratings
• Quantify qualitative data to analyze success factors of webtoons

🚀 Key Features
• Sentiment analysis using Hugging Face (positive/negative classification)
• Popularity prediction models (Random Forest, Regression, LSTM)
• Rating comparison visualizations by genre and weekday
• Prediction accuracy comparison (actual vs. predicted)

💻 Responsibilities & Contributions
• Built and debugged Selenium-based crawler
• Designed FastAPI backend and MongoDB schema
• Built web dashboard UI and implemented chart visualizations
• Designed full data pipeline and handled API-model integration

✅ Outcomes & Retrospective
• Applied multiple AI/ML models in a short time with focused development
• Implemented a fully integrated end-to-end pipeline from crawling to visualization
• Built a solid case of popularity trend analysis using actual webtoon data