A few weeks ago, Max van Duyne was in a bad place mentally. He felt anxious and depressed; laying awake at night criticizing the smallest mistakes and negative personality traits. He wanted to seek professional help, however he moves often and had a bad experience with tele-therapists in the past. That’s when he developed the idea. A self-therapy app, powered by machine learning, that analyzes and displays his thoughts and emotions more clearly than he can internally.

At TreeHacks 2020, he met Kelly Chan, a data science student from UC Davis. She met Max at his challenge and they both dove into the unknown, creating what is now known as Open Book.

What it does

A person navigates to the Open Book and enters a journal entry on the home page. From there, natural language processing scrubs the entry for sentimentality, labelling the document and each sentence as positive, negative, or neutral. This allows the person to see how they are feeling if they don’t fully know themselves.

How we built it

We used Python, HTML/CSS, and JavaScript to build Open Book. The site is primarily built off of Microsoft Azure, using their App Services to host the site and Cognitive Services for sentimentality testing. On top of these digital items, Open Book was built with our collective enjoyment and frustration.

Challenges we ran into

Neither of us are CS students. Max is a Mechanical Engineering student with almost zero prior knowledge of Python. Kelly knew a little Python beforehand, however we were both unprepared for difficulties pertaining to Azure, JavaScript, and HTML to say a few. At almost every step of the way, we found something we didn’t understand and had to learn our way through this hackathon. Even with guidance, our APIs and software were finicky and broke easily. This lead to a fine balance of experimentation and reserve to make sure that we were learning what we had to do, without breaking the entire thing.

Accomplishments that we’re proud of

We were able to create a minimum viable product, which includes web app with hosting, a working journal entry box and functional natural language processing.

What we learned

We learned a lot. We learned about Microsoft Azure, Python, natural language processing, HTML, CSS, Flask apps, JSONs,

What's next for Open Book

Since our submission is only the minimum viable product, we still have a lot to go. Potential next steps for Open Book Include: *Language and keyword detection from Azure Cognitive Services *Timed journal writing *Improved UI/UX *Timestamps and historical data *Timelines *Improved data visualization *Regular expressions, leading to improved pattern matching

Share this project: