Sympthia

Inspiration

The initial anxiety of not knowing whether a symptom is a true emergency or something you can manage at home is a universal stressor. We were inspired by a dual problem: how can we empower individuals to make better-informed decisions about their health, while also providing our public health authorities with the tools to be more proactive instead of reactive? We envisioned a system that could ease the burden on both patients and the healthcare system simultaneously.

What it does

Our application is a two-part public health intelligence platform: For the User: A conversational AI agent that acts as a preliminary symptom navigator. A user can describe their symptoms in natural language, and the agent, following a carefully designed logic tree based on public health guidelines, helps them determine if their situation warrants an immediate trip to the ER or if it can be safely managed at home. For Health Authorities: Every interaction provides valuable, but completely anonymized, data. The type of symptom and the user's general location (e.g., postal code area) are sent to a cloud backend. This data is then aggregated and visualized on a real-time dashboard, allowing public health officials to spot trends, identify potential outbreak hotspots, and allocate resources more effectively.

How we built it

We built this as a full-stack cloud-native application, leveraging the power and scalability of Google Cloud Platform (GCP). User Input (Flutter app) → Cloud Functions → Dialogflow ES → Cloud Functions → Bigquery ↓ Real-time Public Health Dashboard (Looker)

Challenges we ran into

We had two main challenges: a non technical related to ethics and a technical We are not doctors, so creating a reliable triage logic was a significant hurdle. We overcame this by basing our conversational flow building in very clear disclaimers that our app is an informational tool, not a substitute for professional medical advice. On the other side, building an app with such technical complexity in a short period of time was not easy, but we learnt a lot.

Accomplishments that we're proud of

We are proud of building a functional, end to end solution that addresses a real world problem from two different angles. Seeing the first piece of anonymized data travel from the chat agent all the way to a live dot on our dashboard map was a huge moment for us.

What we learned

We learned how to design and train a conversational agent in Dialogflow from scratch. More importantly, we learned how to build a scalable, real-time data pipeline using serverless functions and modern data warehousing tools like BigQuery. On a softer note, we gained a profound appreciation for the complexities of the healthcare space and the critical importance of designing technology with ethics and privacy at its core.

What's next for Sympthia

There are a lot of future developmentsfor the app, but the main are predict the virus propagation behavior, using an AI voice assistant and connecting the app to ER services