The companion you trust, the friend that’s never busy, the teddy bear by your bedside. Hello Teddie!

Inspiration and Research

As engineering and university students, the feeling of depression and desolation is not an unfamiliar feeling. Many occasions throughout the year, we’ve found ourselves descending into a constant state of complaining and doubt. Through talks with fellow classmates, we realize that we were are alone. The fact, however, makes us more isolated as you realize others don't have the mental capacity to comfort you.

Further investment reveals that mental health is not accessible to everyone. Two key factors play a role: money and supply. According to the Office of Adolescent Health, 10% of adolescents lack insurance and have a lack of mental health services due to geographic and social factors. The university of Waterloo, as of March 28, 2018 has 22 full-time equivalent counselling services staff and 2 full-time equivalent psychiatrists serving over 36000 undergrad and grad students. That's a 1500: 1 ratio!


Teddie fills the gap by creating a reliable companion that you can talk to at anytime. We target anyone ranging from occasional anxiety to depression in any age group (even younger people). According to National Institute of Mental Health, over 20% of adolescent teens suffer from mental illness and over half go undiagnosed. This suggests that there is a huge potentially untapped market.

Teddie can also assist with remote therapy. Providing a graph with more detailed information to therapist can assist in their understand of the patient. Hence also reducing the frequency of therapy session leading to more available space.


Teddie functions include a voice feedback system with Amazon Alexa. Using Google Cloud API, we can analyze the text in one’s speech to determine the positivity or negativity through tone and other linguistic features. This data is plotted in a timeline which allows the user to see how they actually been feeling in a period of time. We intently removed the specific memory attached with the sentiment the app is not about nostalgia, especially if its was a bad memory. The action of just talking out loud, proven by research from Michigan State University and University of Michigan, can lead to a reduction in anxiety. Teddie also generates of a chart of 5 emotions: fear, sadness, joy, confidence, anger. To further help with assessment.


Amazon echo (Alexa) connects to AWS to send and receive user input and is sent to google cloud compute engine using Amazon lambda. Google cloud runs backend which is built with python and Django. Backend does natural language processing using two separate APIs: Google Cloud sentiment analysis which returns a score -1 to 1 (negative to positive) and IBM tone analyzer with will return a category of tone. The data is sent to Firebase. Our React front end will live update from the Firebase to produce our chart and trend plot.

Share this project: