Inspiration
Life can’t get busy enough. Whether it’s working long and strenuous hours or balancing extracurriculars, we’ve become so consumed in our everyday lives that we’ve forgotten about the existence of breaks.
After reading the workplace strategies for mental health, we came across a section that described 3 types of breaks, all equally as important. Calming breaks like journaling, and meditation, energizing breaks such as exercise, and relaxation breaks such as trying out soothing breathing techniques. If only there were a way to track the amount of breaks you take…
What it does
That’s where Bliss comes into play. Bliss is a Taipy web app where workers can keep track of the number and type of breaks they take each day. They get to choose from a pre-made list of efficient break exercises recommended by Canada Life and Bliss has its own chatbot, Arcadia, which also pitches suggestions using Cohere.
How we built it
We built the app completely with Taipy front-end and Python back-end. Data is written and read from a CSV file that stores the type of breaks (calming, relaxing, energizing) the user takes as well as the current date. The Pandas library was used to collect data from the files and output them into pretty charts with the handy assistance of Taipy's Charts.
Using Cohere's text-generation API, we were able to build a chatbot capable of generating an unlimited amount of break ideas. By fine-tuning the max-tokens, max word count, and temperature to get helpful recommendations, we tailored the language model to our product.
Challenges we ran into
There are not too many resources for Taipy aside from the official documentation and their set of 5 videos. This meant that our resources were sparse and had to rely mostly on problem-solving using our previous knowledge of web development and the combination of our last brain cells. Some features worked better with markup than with HTML and vice versa, challenging our understanding of Taipy.
Accomplishments that we're proud of
It was the first time we heard about Taipy and we are proud to have picked it up quickly. Also, we decided 30 hours into hacking that AI would be cool to try for the first time.
What we learned
Python can literally do anything and never go to karaoke if you want to save face.
What's next for bliss
We fell short of a full stack application as we decided to not roll with mongoDB and alternatively use CSV files. In addition, we thought of implementing a journal space where a topic generated by AI (relating to mental well-being improvements, or fun ideas like “argue if cereal is a soup”) can be a starting point for users to write their thoughts down.
We also hope to include some sort of live breathing tracker within the app. Users would be able to follow guided intervals for breathing. We were thinking of using some sort of facial detector, to better guide users through these breathing exercises as Taipy had an existing project related to this.
Log in or sign up for Devpost to join the conversation.