Mental health is a major issue especially on college campuses. The two main challenges are diagnosis and treatment.


Existing mental health apps require the use to proactively input their mood, their thoughts, and concerns. With these apps, it's easy to hide their true feelings.

We wanted to find a better solution using machine learning. Mira uses visual emotion detection and sentiment analysis to determine how they're really feeling.

At the same time, we wanted to use an everyday household object to make it accessible to everyone.


Mira focuses on being engaging and keeping track of their emotional state. She allows them to see their emotional state and history, and then analyze why they're feeling that way using the journal.

Technical Details


The user's speech is being heard by the Amazon Alexa, which parses the speech and passes it to a backend server. Alexa listens to the user's descriptions of their day, or if they have anything on their mind, and responds with encouraging responses matching the user's speech.

IBM Watson/Bluemix

The speech from Alexa is being read to IBM Watson which performs sentiment analysis on the speech to see how the user is actively feeling from their text.

Google App Engine

The backend server is being hosted entirely on Google App Engine. This facilitates the connections with the Google Cloud Vision API and makes deployment easier. We also used Google Datastore to store all of the user's journal messages so they can see their past thoughts.

Google Vision Machine Learning

We take photos using a camera built into the mirror. The photos are then sent to the Vision ML API, which finds the user's face and gets the user's emotions from each photo. They're then stored directly into Google Datastore which integrates well with Google App Engine

Data Visualization

Each user can visualize their mental history through a series of graphs. The graphs are each color-coded to certain emotional states (Ex. Red - Anger, Yellow - Joy). They can then follow their emotional states through those time periods and reflect on their actions, or thoughts in the mood journal.

Built With

Share this project: