Data-Driven Blogging Trends Analysis: Sentiment Detection and ML-based Classification with Interactive Dashboards

Analyzing blog content has become essential for comprehending new trends, user interaction, and public mood due to the exponential proliferation of online blogs. This study uses web-scraped data to identify blogs and analyze sentiment using machine learning. RSS feeds from various technology-focused websites were used to gather blog data. Several classification models, such as Naïve Bayes, Support Vector Machine (SVM), and Random Forest, were used to preprocess and analyze the dataset to categorize blogs. Additionally, the lexicon-based VADER and deep learning-based BERT models were used for sentiment analysis. According to the comparing results, BERT had the highest accuracy for sentiment analysis, whereas Naïve Bayes fared better than other models for blog classification. To visualize trends and facilitate real-time blog analysis, a Streamlit dashboard was created. This study provides insights for scholars and industry experts by demonstrating the effectiveness of combining machine learning and natural language processing algorithms for automated blog analysis.