Social Impact BuildFest – EveryLibrary Advocacy Project
Overview
As part of the Social Impact BuildFest, I developed a Python GUI that enhances digital advocacy using sentiment analysis and text-to-speech to drive haptic feedback on DataFeel devices. The tool analyzes text sentence-by-sentence, classifying emotional intensity using rule-based (NLTK, MWEs) and transformer-based (Hugging Face) models.
This project focuses on accessibility by incorporating dyslexia-friendly fonts, UI contrast adjustments, and customizable text settings, ensuring a more inclusive advocacy experience.
Key Features
- Sentiment Analysis Engine: Combines rule-based NLP (NLTK lexicons, MWEs) and transformer-based models (Hugging Face) to classify emotional intensity.
- Haptic Feedback Integration: Converts emotion-based text classification into DataFeel haptic commands.
- Accessibility Enhancements: Supports dyslexia-friendly fonts (OpenDyslexic, Atkinson Hyperlegible), adjustable contrast, and font size options.
- Automation & Visualization: Enables batch processing of text files and provides emotion timeline visualization using matplotlib.
Tools Used
- Programming Languages: Python
- Libraries/Frameworks: PyQt6, NLTK, Hugging Face, Matplotlib, Pyttsx3
- Hardware Integration: DataFeel SDK
Impact
- Improved digital advocacy accessibility, enabling a multi-sensory experience for individuals with visual impairments or learning disabilities.
- Created a user-friendly, customizable interface to enhance text-based emotional engagement.