Inspiration
The inspiration for Traige (TRAiGE) came from observing the critical delays and inefficiencies in emergency response scenarios, particularly within military operations and disaster-hit areas. We were motivated by the potential of wearable technology and advanced analytics to revolutionize the way triage is performed, making it faster, more accurate, and data-driven. The goal was to leverage technology to save lives by ensuring that immediate care is provided to those in dire need, inspired by the ethos of leaving no one behind.
What it does
Traige is an automated triage response system that utilizes wearable technology to monitor vital signs in real-time, coupled with a machine learning model to analyze this data for urgent care needs. It identifies individuals who require immediate medical attention, categorizing them into health statuses: healthy, wounded, critical, or deceased. The system streamlines the decision-making process for medics in the field, enabling them to prioritize care based on real-time data analytics.
How we built it
We built Traige using a combination of FastAPI for the backend to simulate data production from wearables and to serve as the consumer that processes incoming data. The machine learning model was developed using Python's scikit-learn library, trained on historical health data to recognize patterns indicating different health statuses. For the frontend, we developed a Flutter application that displays the prioritized list of individuals based on the real-time analytics processed by our backend. Docker containers were used to encapsulate the producer and consumer components, ensuring scalability and ease of deployment.
Challenges we ran into
It took a lot of adjustments to train an accurate neural network for the features we found. Training was also a problem, luckily we discovered a large data set that encapsulated vital metrics.
Accomplishments that we're proud of
We are particularly proud of developing a functional prototype that can accurately predict urgent care needs based on real-time data. Overcoming the technical challenges to ensure seamless integration between different components of Traige was a significant achievement. Additionally, the potential impact of our project in saving lives and optimizing emergency response efforts gives us a tremendous sense of accomplishment.
What we learned
Throughout the development of Traige, we gained valuable insights into the complexities of real-time data processing and the intricacies of machine learning model development, especially in a healthcare context. We also learned about the importance of user interface design in developing applications for high-stress environments, ensuring that information is presented clearly and concisely to facilitate quick decision-making.
What's next for Traige
Looking ahead, we aim to refine our machine learning model with more comprehensive datasets to enhance its accuracy and predictive capabilities. We also plan to diversify the product to a range of use cases, like first response and incident response systems. Further to this, the app is ready to be scaled - with expansion of more data producers and consumers. We are proud to have build a sustainable real time app and can't wait to go further.
Log in or sign up for Devpost to join the conversation.