Inspiration
make access to everyday therapy faster, cheaper (free), and easier while keeping your identity anonymous.
What it does
hippo checks up on you every day, records how you are feeling in an unintrusive manner throughout the conversation, analyzes the emotions behind your language and patterns in your mood, as well as provides helpful resources to manage your feelings.
How we built it
We designed in Figma to HTML and CSS, machine learning on Kaggle GPU and integrated our components into a Flask full stack app. We used a Chatterbot Python bot for conversational interaction.
Challenges we ran into
Time was the biggest challenge, especially for a machine learning project where just training the model can take up to 5+ hours. In addition, the team was introduced to a completely new programming language and most of us had no experience with making a chatbot. We experimented with various AI chatbot platforms, including Dialogflow, before settling on Chatterbot that was most suitable.
Accomplishments that we're proud of
- Successfully analyzing moods in conversational language
- Creating a working chatbot that can maintain a short conversation
- Coming up with a solution that helps those unable to access a real therapist due to family situation, finances, social stigma. This anonymous, time-efficient tool can help users keep track of the emotions surrounding their thoughts.
- Our neat mascot, hippo!
What we learned
- Configuring AI chatbots and integrating them into our web applications
- Designing intuitive applications with Figma, HTML and CSS
- Building machine learning models to classify emotion based on text
What's next for hippo
Product market fit! Getting this out there to users to use, get feedback, improve and iterate. We want to develop our mood tracking features to parse diary entries and provide our users with live therapists for when they need expert attention.
Built With
- chatterbot
- css3
- figma
- flask
- html5
- keras
- machine-learning
- natural-language-processing
- python
- tensorflow


Log in or sign up for Devpost to join the conversation.