Decision Bot

The Hidden Psychology Behind Big Decisions

Everybody has come across what seems like a life changing decision and thought about it for days, weeks, months, or even years. Why are these decisions so hard to make? The Decision Bot that we created helps you break down the decision by asking you ten discriminating questions that help you determine whether or not to go through with it.

How we built it

With the Google Assistant, the human interface is conversational so we used Dialogflow to design the UI. To fulfill the user actions, we used a serverless architecture with a Google Cloud Function to send their voice input through our machine learning pipeline to determine whether or not you're ready to make the decision.

Challenges we ran into

Our biggest challenge was setting up the pipeline for data analysis. Also, we had to learn NLP principles.

Accomplishments that we're proud of

Our biggest accomplishment was using natural language processing in the backend to classify user sentiment on a normalized scale.

What we learned

This is our first hackathon so we learned how to manage a project with a team on such a short deadline. We also developed a further understanding of JavaScript, machine learning, how to use an IDE called WebStorm, and the inner workings of Git.

What's next for Decision Bot

We want to integrate our Decision Bot with other APIs such as the users bank account to better determine whether or not the decision is good for the user.

Built With

• firebase
• google-action
• google-machine-learning
• javascript
• node.js