JournAI

• 

  This is the way we show users their 7 core emotions and their relative frequency

Inspiration

We recognized the popularity of Spotify wrapped. It provides an interesting breakdown of people's personalities in terms of their music taste! We thought it would be great to follow that idea, but in a much more personal way. We also identified the problem that people hardly every get the chance to reflect on the emotions that they've felt in the past usually its because they're so busy that they move on and forget. But if there was some way that you could visualize the emotions that you've felt over the past it would allow for better self-reflection. Think of it like if you could analyze a mood ring that you've been wearing.

What it does

Our app is built to feel like second nature for people who want to start or already are blogging and journaling regularly. With the help of natural language processing our app classifies entries into 7 core emotions. These core emotions are then categorized into different charts to help the visualize their modes of speaking. Doing so creates sort of a profile of their emotions.

How we built it

We used Next.js and Firebase as our tech stack to combine quick development with even quicker requests. For analyzing the sentiments, we employed the use of a pre-trained Natural Language Processing model, who's API return us the emotions of a particular entry.

Challenges we ran into

We started the Hackathon using a library in Javascript with very poor documentation and outdated examples. We sunk about a third of our time struggling to make the library work, before realizing we needed a different solution. Once we transitioned to the new library, we were on a better pace, but were up against a time crunch to deliver our submission on time.

Accomplishments that we're proud of

We developed a fully functioning web app that acts how we envisioned it would! We're handling data in an efficient and scalable manner and we're very proud to have employed the use of Natural Language Processing.

What we learned

We learned how to trouble shoot and delegate tasks. We frequently ran into merge conflicts, and misleading documentation. We had to come together to find solutions and not waste time. We also learned that things do not always go to plan, and that being flexible is important.

What's next for JournAI

We would like to fully support mobile as well as develop a more in depth visualization of the data. We plan to make interactive 3D models of the data as well. Once satisfied with the data visualization end, we will tackle building and training out own NLP model.

Built With

• chart.js
• firebase
• huggingfaceapi
• next.js
• tailwind.css