DiseasePred Insights

• 

  DiseasePred Insights

Inspiration: The rising cost of healthcare, especially for early diagnosis, inspired me to create DiseasePred Insights. In many parts of the world, people hesitate to seek medical advice due to the expensive and time-consuming process of disease diagnosis. I wanted to build a tool that empowers users to understand their symptoms and take early action for free and from anywhere, right from their browser.

What It Does: DiseasePred Insights is a web application that allows users to: Select their symptoms using a simple, user-friendly interface. Predict the most likely disease using a pre-trained machine learning model. Instantly get a summarized explanation of the disease from Wikipedia. Search for nearby hospitals or clinics for the predicted condition using DuckDuckGo, with links to book appointments.

How I Built It: Frontend: Streamlit, chosen for its speed and interactivity. Model: Integrated a pre-trained scikit-learn model from AWeirdDev/human-disease-prediction. Information: Used the Wikipedia API to fetch summaries of predicted diseases. Hospital Search: Leveraged the DuckDuckGo Search API to find hospitals near the user's location. Deployment: Hosted the app on Streamlit Cloud for easy public access.

Challenges I Ran Into: One of the major challenges I faced was selecting a suitable model that balanced both prediction speed and accuracy. Since medical predictions demand reliability, I had to carefully assess models and validate their performance within the constraints of a real-time web app. Additionally, integrating external services like Wikipedia and DuckDuckGo added complexity in terms of response time and handling edge cases, such as missing pages or ambiguous search terms.

Accomplishments That I'm Proud Of: I’m proud that DiseasePred Insights is not just a working prediction tool, but also a practical health assistant. Additonally, I am also proud that it is able to fetch disease descriptions from Wikipedia and recommend nearby hospitals for treatment through DuckDuckGo made the project feel complete and impactful. .

What I Learned: Through developing this project, I learned how to efficiently use third-party APIs to extend the capabilities of a web app. I gained experience using the Wikipedia library for fetching useful summaries, and in leveraging DuckDuckGo’s search engine to return location-based results. I also improved my understanding of model integration and caching in Streamlit for better performance.

What’s Next for DiseasePred Insights: In the future, I plan to integrate a more medically accurate model, potentially using neural networks or models trained on real-world clinical data. A major improvement will be the inclusion of login and sign-up functionality, allowing users to track their health history, receive personalized suggestions, and securely store previous results. Additionally, I hope to even integrate with wearable health data to offer deeper insights and early warnings.

Built With

• api
• duckduckgo
• huggingface
• python
• scikit-learn
• search
• streamlit
• wikipedia