Inspiration
The inspiration for this project is simple: save more lives. Every person on this team has known someone who at one point had cancer or another genetic disease. Thus, we have all been exposed to the harsh reality many individuals face when having to live with such diseases. We also know that the earlier a disease is detected (or the earlier someone knows the risk they have of facing a disease), the better the chance they have of surviving. Thus, with Falcon, we set out to develop a platform that detects a patient's risk of facing a disease early on and successfully makes life-saving recommendations based on a patient's data.
What it does
Project Falcon is an end-to-end progressive web app built to analyze the cancer risk category of individuals so that geneticists can make better-informed decisions. The application currently allows counselors to view and sort through thousands of patients using a large number of data points. Users of the application are also able to quickly and conveniently enter new patient information to determine their risk of developing cancer-based on a variety of factors, including their family history. This information is processed using our developed algorithm to make a smart decision about their risk category.
How we built it
There are two layers to the Falcon application. The first layer is the web app, which is where all user interaction occurs. This web app was built using modern technologies, including Vue.js, Bootstrap, and TypeScript. The second layer is the backend application, which is responsible for data storage and retrieval, as well as for the determination of patient risk. The backend application was written in the Go programming language, using an SQL-based ORM for data storage. These two layers combine to provide the functionality needed by our users.
Challenges we ran into
This hackathon project was especially challenging since all members were collaborating remotely, and many of us were in the middle of finals week for school. Some of the other challenges we faced included familiarizing ourselves with the tech stack we chose and working with a data set that contains thousands of records. Also, with none of us having experience in artificial intelligence or machine learning before this hackathon, implementing an AI solution in our code was also new territory to us. However, we didn't let these obstacles deter us, working extremely hard to accomplish all project objectives.
Accomplishments that we're proud of
We are proud of this team's ability to overcome the hurdles associated with distance-based collaboration by maintaining consistent communication and establishing our individual roles early on. By delegating roles to each team member, we were able to play to each team member's strengths and in turn, create an optimal solution to the hackathon challenge we selected. Although it was not easy to fit this hackathon into our busy week, we were committed to the project and came out with a solution that addresses the challenge statement.
What we learned
We learned many things in developing the Falcon platform. Before this competition, nobody on this team had much experience in AI or Machine Learning, but through hard work and tenacity, we were able to create a solution that addresses the challenge description. In addition to this, we learned how to design software solutions that process large data sets with thousands of data records. We also learned how to create efficient and effective models that categorize data and sort it in a manner that allows for accurate recommendations. We even learned how to take this data model and create a visual interface that makes it easy for clients (in this case healthcare providers) to interpret the findings of our data model. Lastly, we learned how to develop a full-scale project in a virtual environment, relying on effective communication and time management to make this a reality.
What's next for Falcon
In Falcon 2.0, we would enjoy the opportunity to better improve on our decision-tree AI to make it more accurate at determining someone's risk of developing cancer. Unfortunately in this competition, we did not have the necessary time to implement any sort of machine learning or neural networks, which is something we would love to do. This also means that users are currently not able to provide corrective feedback to our AI. In addition to this, we would love to create iOS and Android versions of this app to increase the accessibility of the Falcon platform and thus save more lives.
Log in or sign up for Devpost to join the conversation.