NeuroVoice- Parkinson's Health Assistant

💡 Inspiration

Parkinson’s affects over 10 million people worldwide, yet early detection remains difficult. We wanted to create a non-invasive, accessible tool that uses something everyone has — their 🎤 voice — to detect early signs and raise awareness.

⚙️ What it does

NeuroVoice analyzes subtle voice frequency and tone variations to predict the likelihood of Parkinson’s disease. It provides clear, explainable AI results 🧠 and trusted medical resources for next steps.

🏗️ How we built it

We trained a Random Forest model on the Parkinson’s Disease Detection Dataset (Kaggle) using features like jitter, shimmer, and harmonic ratios. The app was built with Streamlit, offering real-time predictions and feature importance visualizations 📊.

🚧 Challenges we ran into

Managing inconsistent feature columns across datasets and ensuring accurate feature scaling. We also worked to balance scientific precision ⚖️ with a friendly, empathetic user interface that feels approachable.

🏆 Accomplishments that we're proud of

Creating an AI tool that detects Parkinson’s risk early with strong accuracy, designing a beautiful and accessible interface 💻, and implementing explainable AI visualizations that promote user trust and understanding.

📚 What we learned

We learned how to handle biomedical data, apply explainable machine learning, and design AI systems that make health insights human-friendly 💬.

🚀 What's next for NeuroVoice – Parkinson's Health Assistant

Our main goal is to collect more data and strengthen the model further. Additionally, we plan to integrate live voice recording analysis, multilingual support, and a mobile-friendly version 🌍. Collaborating with healthcare professionals for real-world validation is our next major milestone 🩺.

Built With

• kaggle
• python
• random-forest-classifier
• streamlit