Granola

Inspiration

The inspiration for this bot was when I (Matthew) was in Band Class this Friday. My band teacher had a survey of the challenges of this pandemic. Most of the people in the band said that their mental health was the biggest problem that they have faced during the pandemic because they had no one to talk to. Because of that, I got the idea that we could potentially make a chatbot for people who do not have easy access to therapists in the world and do not have money.

What it does

This chatbot is a conversational chatbot that gives you advice on what to do when you are feeling a certain emotion. The emotions which we have made for the chatbot in the 24 hours given are angry, sad, and happy.

How we built it

We used Dialogflow which is an API that allows you to create and design a chatbot by showing the bot certain keywords which it has to recognize when a user says something. When the bot recognizes the keyword, it will determine which intent the user has and then respond based on the intent.

Challenges we ran into

We had several challenges during the hackathon. Our first challenge during the hackathon was that we attempted to use repl.it to code our chatbot but when we ran the chatbot, repl.it crashed. Then, we noticed that repl.it wasn’t a good platform to code a chatbot so then we did some research and decided to use Dialogflow which is made from Google Cloud. We had to learn the basics of Dialogflow and natural language processing so that we could teach the bot about synonyms of certain keywords so that the bot could have a simple conversation with you.

Accomplishments that we're proud of

We are proud of being able to make a chatbot using an API that we have never used before in such a short time span. We also had to overcome the problem of our idea not going how we originally planned it to go, but still being able to find a solution that worked for our idea.

What we learned

We learned that making a chatbot can be very difficult. Originally, we wanted to make the chatbot using repl.it, but with the skills we had, we soon came to realize that we would not be able to create this bot on repl.it. So, we did some research and came up with a solution that worked; and that was to use Dialogflow to create this bot. In Dialogflow, we learned how the bot can recognize specific keywords and tell what emotion the user is feeling. We did that by learning what intentions are which is when the bot recognizes the intent of the user and how they are feeling during the conversation. The bot can find the intent or emotion of the user by finding certain keywords like “sad”, “happy”, or “angry”. Finally, we learned how to “train” the bot to use synonyms like “unhappy” and understand that the word "unhappy" relates to the emotion of "sadness".

What's next for Granola

Our next plan for the Granola Bot is to make the bot be able to save the user's past conversations with the bot so the bot can tailor its responses towards the user. Also, we want to make the bot to be able to have more complex conversations by teaching the bot to understand more keywords or sentences that users may input.