PathoBot (Early Disease Detection)

• 

  Architectural Design of Project

Inspiration

This project was inspired by the need to make health information more accessible and understandable. I wanted to create a tool where users can input their symptoms and quickly learn about possible diseases in simple language.

What it does

PathoBot helps users identify possible diseases based on the symptoms they select. It predicts the top likely conditions using a Random Forest model and provides simple, easy-to-understand explanations about each disease, including severity and recommended actions.

How I built it

I collected a disease-symptom dataset, cleaned and preprocessed it, and trained a Random Forest model to predict diseases. I used FastAPI to build the web app and integrated Google’s GenAI API to generate user-friendly disease summaries.

Challenges I ran into

One main challenge was finishing a consise version of my idea on time as it uses Gemini API integrations, Model Trainings, and a huge dataset. Handling rare diseases in the dataset orignally gave lower accuracy scores in which I had to compensate with extra trainig time. Another challenge was generating explanations that are both correct and easy for users to understand using the Gemini API which led to several traceback errors.

Accomplishments that we're proud of

I am proud of successfully building a working prototype that combines machine learning and AI-generated content to make complex medical information accessible. Users can now get predictions and explanations in a single interface.

What we learned

I gained experience in data preprocessing, model training, API integration, and building a full-stack AI-powered web application. I also learned the importance of user-friendly explanations in healthcare applications.

What's next for PathoBot (Early Disease Detection)

I plan to expand the dataset for more diseases, deploy the app for wider public use, and developing a mobile version of this application for easier and wider access. I also plan to increase the accuracy to make sure PathoBot is reliable for billions of people.

Built With

• fastapi
• geminiapi
• python
• randomforest