Facets Wellness App

Main Dashboard
Login
Indentify Your Feelings: Sentiment Analysis Results
Indentify Your Values
Select an LLM of preference
AI Assistant ask user about their day
Example with positive emotions
After user input, AI assistant provides motion mapping.

Inspiration

Nonviolent communication uses the ability to speak about ones emotion to resolve conflicts and communicate better with others. The ability to about what one feels helps you regulate your emotions, have self awareness, better mental health and better connections.

See resources: https://hbr.org/2016/11/3-ways-to-better-understand-your-emotions https://www.cnvc.org/

What it does

Facets is a mental wellness app that has numerous activities to help you better understand yourself and gain better self awareness. For this hackathon we focused on building the 'Identify Your Feelings' activity.

Having personally used the list of feelings identified by Marshal Rosenberg in nonviolent communication, I was able to self soothe and calm myself down during turmoiled time. Studies have shown just being able to label your emotions can help decrease stress.

The 'Identify Your Feelings' activity first describes the importance of identifying ones emotions. It then invites the user to journal about the situation that resulted in the feeling of emotions.

Using Non Violent Communication's Feelings List, the user is able to select and identify the possible emotions they're feeling based on whether there needs are being met or unmet. Using drop down and emotions we have made this a more engaging and fun experience.

Lastly we use sentiment analysis to help users become even more aware of emotions they expressed via their journalling. Further deepening and developing self awareness by asking them what they would have liked to happen and what they would have done instead.

The user is able to download their journal into and html format. Many diary and journaling apps have disgruntle users when their data becomes lost or incompatible over time. While other journaling apps allow users to downloads their journals, they are in a hard to use/read json format. By using TinyMCE we were able to easily save their journal entry into html. Making it easy for the user to download, save, reopen and read their entries.

We explored also a chat version of the activity.

How we built it

We used next js, flask and react as the base of the project. For journalling we used tinyMCE, and enabled to user to download html of their journal entry. For the sentimen analysis we used a model created by OpenAI in Clarifai. We then used Langchain with Clarifai to create the output of our sentiment analysis based on the feelings we fed it based on a non violent communications feelings list.

Challenges we ran into

Realizing the API key was provided in the code for TinyMCE. Also getting and open AI API key required spending money. Understanding how Clarifai used different model libraries, workflows and datasets in their interface, and then figuring out which was the authorID, API keys... took time and was confusing as there were too many keys. Understanding input and output, then understanding how to use the workflow took a lot of time. We first mocked up Clarifai using Streamlit but had a hard time converting it back into full python to pass via a react API call. It was our first time using Langchain, we struggled understand how to preset some instructional history via code. We tried building a login using email and password, and took some time to understand access tokens and refresh tokens, but were unable to complete usage of it. We had a simple access token login mostly built. We realized after storing the users' journal into google drive would required using google login. We decided to focus on completing the journalling activity instead.

Accomplishments that we're proud of

We learned a lot. It was the first time use react with a python backend via nextjs/flask. The setup made it easy to get our project started. We also used tailwind and daisyui to ensure our visuals were fast and easy to build up. We used two products from the challenges: TinyMCE and Clarifai We built a piece of our wellness app that we would actually use in our final product.

What we learned

nextjs and flask langchain got slightly better at python first time using Daisy UI

What's next for Facets Wellness App

We will be continuing to build out modules based on psychology, mental wellness, relationship, neuroscience and relationship research. We are currently working with a therapist to build out Acceptance Commitment Therapy modules. Stay tuned for our product at https://facets.one. While we are wellness focused, we are currently targeting singles who are dating since divorce rates are at 50%, people are lonelier than ever, people are burned out from online dating, and relationship issues are the root to family problems and everyone's childhood formation.