Mental Note

Inspiration

Our team was heavily inspired by the many workshops we were able to attend during our incredible time at Hack the MIST! As a result, in our project can be seen many of the different tools and technologies seen in the workshop.

What it does

Mental Note is a sentiment journal that acts as a personal mood tracker. Users write small notes about their day and receive a personalized themed image generated based on the overall sentiment of their input. After the use of Mental Note for an extended period of time, the user is able to view a colourful mosaic of their mood over time.

How we built it

The user gives input about their day into Mental Note. Sentiment analysis is applied to the user input using HuggingFace to detect the user emotion. Using OpenAI's API for text to image generation and neural style transfer learned during Hack the Mist workshops, an image is automatically generated that incorporates aspects of both the content and the mood of the user's input. This image is then put into a calendar to create a visualization of the user's mood over time.

Challenges we ran into

A general challenge our team faced was general unfamiliarity with the technology. With no previous ML/AI experience, and some basic python syntax knowledge in hand, we came into Hack the MIST with the intention of learning as much as we can and having as much fun as we could while exploring this exciting field. Given our lack of experience, this presented a lot of challenges in navigating Flask, HuggingFace, neural style transfer, and also implementing VoiceFlow. Another key challenge was trying to connect all of these moving pieces into a cohesive app.

Accomplishments that we're proud of

Our team is proud of the mindset we were able to come into this hackathon with. We refused to revert to the tech stacks we were comfortable and familiar with and continue treading onward into the unfamiliarity to learn as much as we could. In the end, we have come out of this hackathon having attended all of the beginner workshops and learning a lot along the way.

What we learned

The huge perk of coming into a hackathon focused on an area of tech we were unfamiliar with is that we were able to learn a lot of new knowledge! We joined every workshop we could without regrets (and installed a lot of packages). We were able to have a lot of fun immersing ourselves in the world of ML/AI for the first time with Hack the MIST, and plan on learning even more in our future endeavours!

What's next for Mental Note

Our team struggled with connecting all of the different parts of our application, with all of the technology being unfamiliar. Aside from improving Mental Note to perform, we would love to be able to implement a speech to text function, in which we can perform sentiment analysis on the audio as well, to generate more accurate predictions of user emotion. We would also like to give more creative freedom to users to potentially upload their own style images to be transformed by the mood detected by Mental Note. These features would add more user flexibility and customization.