Healthcare should be accessible to everyone, at all times. However, due to issues such as lack of infrastructure and unseen rise of epidemics, such a necessity it not available to everyone. Additionally, with long lines at hospitals, a simple cold can develop into a aggravated flu or make oneself vulnerable to other types of diseases or epidemics. Thus, with the goal of helping people to understand the severity of their situation, we wanted to develop a personal assistant that would guide the user with remedies for proper self-diagnosis and more importantly alert them when it is right to take medical assistance.

What it does

Currently, DocInPoc can be accessed through Google Assistant and it simulates two demonstrable situations where the assistant can be useful. One situation involves a user in Canada describing common cold/ flu symptoms, whereas the other scenario is located in Brazil and helps the user determine if they are a victim to the Chagas infection. These situations are just two examples that show how the assistant can be used globally to identify and differentiate between common illnesses and viral infections specific to a location.

How we built it

We used VoiceFlow and DialogFlow to create the dialog between the user and the embedded assistant. Additionally, we used javascript to sort out dialog inputs, transfer symptom keywords via arrays between intents and display messages back to the user.

Challenges we ran into

There were couple of challenges that we ran into while developing our assistant. One of the included trying to establish a connection between VoiceFlow and DialogFlow to use trained entities and basically communicate data from one platform to another. Furthermore, the only language that was supported by VoiceFlow was Javascript. Thus, trying to understand the extent of the language and how it could be used to configure behaviors in VoiceFlow was another challenge especially due to the lack of knowledge of this programming language within the team.

Accomplishments that we're proud of

Learning how to use new tools and languages such as:

  • VoiceFlow
  • DialogFlow
  • Javascript

What we learned

Implement voiceflow and DialogFlow into the Google Assistant API, as well as how to utilize Javascript to glue the project together/operate and manipulate user input.

What's next for DocInPoc

DocInPoc has a lot of potential to grow in the future. Given more time, we would like to integrate Machine Learning and Natural Language Understanding to expand the reach of the assistant to all countries around the world. Machine Learning would be used to improve the conversation between the user and the assistant so that it is more personal to the user. Natural Language Understanding could be used to develop the assistant so that it can understand and communicate back to the user in more than one language. One of the limitations of our team at the time of developing this assistant was the lack of knowledge of Machine Learning.

Built With

Share this project: