SicklySage

Diagnosed with cancer by the Internet again? Introducing SicklySage – your AI-powered health assistant. We help patients get early diagnoses by analyzing their symptoms and medical history.

Comment

Inspiration

Diagnosis requires doctors’ expertise and may not be accessible to everyone from an early stage, which may cause delayed diagnosis of serious diseases. SicklySage is inspired by the need of making healthcare more accessible and personalized.

What it does

Our project uses Google Gemini to help patients diagnose their symptoms, considering their vaccine and illness histories.

How we built it

We started by deploying a cluster and building the data and database on MongoDB Atlas. Then, we looked into Streamlit and the Gemini API to create a website where users could ask about their symptoms. We also created a login page to store users' vaccine and illness histories. After that, we deployed our website to Streamlit Cloud!

Challenges we ran into

Fetching Data from MongoDB Atlas in combination with managing streamlit’s sessionstate variables took us some time to figure out how to retrieve and present data in the form we wanted. For our login page, we planned to use Streamlit-Authenticator. However, we couldn't figure out how to add the logins in our database to the authenticator, so we had to change our plan. It was also particularly challenging to build an entire project in such a short time. We had ideas but didn't have enough time to complete all of them.

Accomplishments that we're proud of

We are proud of using Gemini so it considers our data and database when answering questions. We’re also proud of deploying our application of Streamlit Community Cloud.

What we learned

We learned about MongoDB, Streamlit, and Gen AI, which we've heard of but have little hands-on experience with. Working on this project, we have had the chance to not only dive into each technology and consider how to incorporate it but also brainstorm and plan a project from scratch.

What's next for SicklySage

We plan to add the doctor role to our project. After Gemini diagnoses a patient's symptoms, it will save the diagnosis. If the patient goes to a doctor afterward, the doctor can update or confirm the diagnosis. With this data, we can fine-tune the model. We also plan to add the nearest clinic/hospital locations, so that after Gemini diagnoses, it will show those locations and ask if the patients would like to book an appointment.

Built With

Share this project:

Updates