Origin
With how Japan has a rigid medical facility system with specialized clinics or hospitals that have limited opening hours, language capabilities and confusing websites to navigate, there are even services where you call an intermediary on the phone that sorts out the best hospital options to consider. Foreigners, for example, are much more limited and cannot access all the information. With the Medi chatbot, it aims to add convenience in narrowing appropriate facilities and also empower the user by providing options to his preference.
What it does
Medi is a quirky chatbot that asks you brief questions before compiling it in order to recommend a medical facility with factors such as symptoms, location, and medical history.
How we built it
Utilizing Node JS as a server to accommodate various third-party APIs, we created a chatbot based on Facebook's messenger platform and the Graph API and programmed it with the _ api medic _ API (courtesy of Rakuten) to analyze symptoms and ask specific medical-related questions. Afterward, the user location is then acquired and passed through the Google Maps API to list the recommended facilities based on personal preferences such as transportation price or access ease.
Challenges we ran into
It took us over 12 hours to get the bot working online with the servers. We started out with the Azure web services but the latest bot builds were extremely buggy which forced us to move to Messenger that was later coupled with Facebook's natural language processing for easy input. Additionally, the team's inexperience with chatbot services, in general, slowed us down, but alongside that ended up making this a fantastic learning experience.
Accomplishments that we're proud of
We divided the team into two to work on entirely different API support and were able to integrate them with the messenger platform built separately. It is the first time for the devs to work exclusively with chatbot systems so it felt amazing to finally get it running the way we wanted.
What we learned
Surprisingly, we learned to take the importance of breaks. Especially after the 10 hours getting stuck, we decided to enforce outdoor breaks which we all reflected and that I believe helped our efficiency afterward.
What's next for Medi
Due to lack of stability in the platform we switched from Microsoft Azure to Facebook, we also wanted to have a floating avatar of our mascot, Medi, some sort of AR (Augmented Reality) support for aesthetic purposes. We also wanted to be able to provide a system that can display recommended doctors with their schedules provided, and continue working closely with hospitals all over the country to make the healthcare system a bit better for everyone.
Built With
- api
- api-medic
- apis
- facebook-chat
- google-maps
- hardware
- hosts
- languages
- libraries
- node.js
- ui-kits

Log in or sign up for Devpost to join the conversation.