Inspiration

Our inspiration for this project was derived from the struggles hospitals are faced with currently during the COVID-19 pandemic in obtaining essential medical supplies and PPE. This is a key socioeconomic issue currently impacting our community: hospitals are unable to keep track of the medical supplies they have in these frantic and unpredictable conditions, and they're unable to safely treat patients affected by COVID-19. Both hospital workers and patients are at risk when doctors do not use proper medical supplies, and therefore, a large loss of life occurs which affects both our society and economy. We know doctors in our own social circles and we have heard of the countless stories of nurses and doctors without PPE or specific medical supplies aiding patients in critical condition. We believe that through an easy-to-use application like SupplyChat, hospitals will be able to keep track of their inventory information in times when everything is in disarray such as during the COVID-19 pandemic, so that they know when to restock supplies and how many patients they can safely attend to.

What it does

Our chatbot is based on a deep learning algorithm (natural language processing) and it learns from inventory information provided by the user, which in this case would be a hospital worker. After the chatbot is trained in less than a minute, the user can input the names of specific supplies to learn the available quantity of the supply. After gaining this information, doctors can make critical decisions as to how to strategically attend to patients with the given number of medical supplies that they have.

How we built it

We used frameworks such as Tensorflow-based Keras and NLTK to develop a natural language processing deep learning algorithm that would allow our chatbot to provide dynamic responses given changing conditions at hospitals during COVID-19. We then integrated this chatbot within a web application using Flask, HTML, CSS, and JavaScript to provide access to users across all platforms. We developed a clean, minimalistic website and added authentication/login services to allow the user to save their progress with their chatbot. In the demo, the web app is run on the local machine, but there is also a website to access it.

Challenges we ran into

We mainly ran into issues while integrating our Python backend with our front end. After some research on various frameworks, we integrated our Python backend and front end using Flask and we were able to deploy a web application for SupplyChat. Our team did not have any experience with using Flask, but after some arduous work, we were able to successfully connect our Python backend with our front end.

Accomplishments that we're proud of

We're very proud of how the web application came out. Even though we had zero experience with Flask, we were able to develop a beautiful front end for our chatbot while also successfully integrating our Python backend. Additionally, we are very happy with how we were able to integrate Firebase for database and for authentication to our application to make it more user-friendly and professional.

What we learned

We mainly learned about the power of frameworks such as Flask and how to use them. Additionally, we learned a lot about deep learning and how to develop neural networks from scratch using libraries such as Keras and NLTK. Finally, we also were able to better our front end development skills through the process of perfecting our HTML and JavaScript code.

What's next for SupplyChat

We hope to provide our software to nearby hospitals in the Bay Area for testing. We want to see if SupplyChat is able to optimize the search time for doctors in gaining information for essential medical supplies and PPE. We believe that with this rapid way to learn about the inventory information, doctors and hospital workers will be able to assess their hospital's capacity in treating patients efficiently and safely. If our product is successful in local hospitals, we plan to not only spread it across several California hospitals but also to spread it to the general community by creating new chatbot configurations. These configurations would allow the user to choose which version of SupplyChat they would like to use, either the community or medical version. With this expansion of SupplyChat, we would be able to impact a wider range of people, ultimately helping them in these unprecedented times.

Share this project:

Updates