About This Application

I’m excited to introduce a simple, lightweight web application for analyzing the sentiment of text. This tool is perfect for quickly checking the tone of emails, passages, or paragraphs without needing heavy or expensive AI models.

As of July 2025, the application offers two main features:

• Sentiment Analysis: Determine whether a piece of text is positive or negative using a lexicon-based approach with popular lexicons like VADER and AFINN. This approach is fast, reliable, and doesn’t require any training.
• Formality Analysis: Assess the formality of a text and see which words contribute most to its formality score. Words are highlighted with colors that reflect their influence. The formality score is calculated using a bidirectional LSTM model trained on the ENRON email dataset. Three independent reviewers manually labeled the training data to ensure reliable results.

This application demonstrates that even simple, lightweight models can deliver meaningful insights for everyday text analysis, without the need for complex, resource-intensive AI systems.