Spamshield Sentiment Analysis on Youtube Comments

The exponential growth of social media platforms has resulted in an overwhelming volume of user-generated textual content, making effective content moderation and opinion analysis increasingly challenging. YouTube, as one of the most popular video-sharing platforms, receives millions of comments daily, which include genuine feedback as well as spam, promotional messages, and emotionally charged content. Manual analysis of such large-scale data is inefficient, time-consuming, and prone to inconsistencies. Therefore, there is a growing need for automated systems capable of analyzing user comments and extracting meaningful insights in real time. This paper presents SpamShield, a web-based automated system designed to analyze YouTube comments using Natural Language Processing (NLP) techniques. The proposed system retrieves real-time comments from YouTube videos using the YouTube Data API and performs comprehensive text preprocessing to remove noise and normalize the data. Preprocessing steps include text normalization, removal of special characters and URLs, tokenization, and stop-word elimination, ensuring that the comments are suitable for reliable analysis. The sentiment analysis process is implemented using the Natural Language Toolkit (NLTK) along with the VADER (Valence Aware Dictionary and sEntiment Reasoner) lexicon, hich is specifically optimized for analyzing social media text. Each comment is evaluated based on lexical and contextual features, and the sentiment polarity is classified into positive, negative, or neutral categories. The system further aggregates sentiment results to generate comprehensive sentiment reports that provide a clear overview of audience opinions and engagement patterns. Experimental evaluation demonstrates that the proposed system effectively analyzes informal and unstructured social media text, offering reliable sentiment classification and intuitive visualization of results. The modular and scalable architecture of SpamShield enables efficient processing of large volumes of comments and supports real-time analysis. The proposed approach provides a practical solution for content creators, marketers, and researchers to understand audience sentiment, enhance user engagement, and support data-driven decision- making in social media environments.