Inspiration

Multiple tragic events in the University of Waterloo community have sparked much debate around the issues of mental health and well-being. In light of these events, we decided to create a chat bot which will talk to students, and provide them with the help and resources that they need.

What it does

WaterlooTalks is a chat bot that provides mental health resources and support for students who seek these things. WaterlooTalks will ask users how they are doing, and respond appropriately to their response by suggesting relevant hotlines, counselling services, or simply providing some encouragement for when one is feeling down. Integrated into Facebook Messenger, WaterlooTalks is easily accessible to users and responds instantly to messages.

How we built it

WaterlooTalks is built on Microsoft’s Azure platform, utilizing Azure Bot Services along with some useful API’s such as LUIS, Sentiment Analysis, and Bing Spell Check. The backend is programmed in Azure Bot Services using Javascript and Node.js, while the front-end website is created using HTML, CSS, and JavaScript.

Challenges we ran into

One of the main challenges we faced throughout the project was getting used to the Azure platform, something that none of our group members had used before. Due to our inexperience, it took a lot of research to just program simple functions. In particular, getting the text string that triggers the appropriate intent took a long time to figure out, as well as making the HTTP request to retrieve the data from the Sentiment Analysis API.

Accomplishments that we're proud of

Something that we are proud of is that we were able to take a platform (Microsoft Azure) and language (Node.js) that none of us were familiar with, and learn enough in a short period of time to make a functioning chat bot. Although we did not successfully implement everything that we intended on doing, we were still able to incorporate a variety of different features of the Azure platform

What we learned

Throughout this hackathon, we were able to learn a lot about many different features of the Microsoft Azure platform, from using Bot Services to integrating API’s, as well as deploying the bot to Facebook Messenger.

What's next for WaterlooTalks

We would definitely like to create more emotions/feelings for users to choose, so that we can cater to more people. We would also like to implement the Cognitive Services better, by improving the Sentiment Recognition implementation in our program. Lastly, we would like to add a wider variety of responses, so that the chat bot does not get boring to use for users.

Built With

  • azure
  • microsoft-bot-framework
  • microsoft-cognitive-services
  • node.js
Share this project:

Updates