Sentiment Analysis of Online Reviews

I worked on a project to help a client analyze customer opinions about their products by examining online reviews. Using Python, Selenium and BeautifulSoup, I scraped reviews from multiple websites to build the initial dataset. The text was cleaned and preprocessed using NLTK, and sentiment was categorized as positive, negative, or neutral using a sentiment analysis tool adapted for Portuguese.
To make the insights accessible, I built an interactive dashboard with Streamlit. The dashboard displayed sentiment breakdowns, key themes, and allowed users to filter data dynamically. I automated the scraping process with a scheduled task that ran daily, updating the dataset, preprocessing pipeline, and dashboard database to ensure near real-time insights.