Have you ever tried keeping a diary, but ended up abandoning it in a few days? You're not alone. We created feels to automatically fill up your journal from events and feelings from your day.

What it does

feels uses machine learning to analyze health metrics, calendar events, and location metadata to create automatic suggestions for when you journal. In addition, feels enhances your personal growth by allowing you to connect with other anonymous users and share experiences with them. We use sentiment analysis to match you with users who have similar emotions as you. All of these techniques are scientifically recommended methods of maintaining emotional well-being. Users can undergo cathartic experiences through self-reflection and sharing.

How we built it

feels was built as a progressive web application, which allows it to provide a rich experience on all platforms. We used node.js and sqlite to build the backend and React for the frontend. Our sentiment analysis was implemented using Azure's text analysis API. Simple metrics such as heartrate are currently used to evaluate a user's mood, but we also plan to examine accelerometer data to detect shakiness.

Challenges we ran into

Debugging asynchronous calls in JavaScript, integrating multiple different APIs with each other, and creating a system that maintains user privacy were our biggest challenges.

Accomplishments that we're proud of

feels solves a problem that all of our team members and many peers have had in the past. We are also proud to have created something that reduces the friction with accessing a form of cognitive-behavioral therapy (CBT) and help people manage their anxiety and depression.

What we learned

We were all pushed out of our comfort zones to learn more about full-stack development, design, and ways to utilize Azure.

What's next for feels

We plan to enhance the diversity of metrics used for mood detection and create a more seamless UX.

Share this project: