Sully

A chatbot that supports local businesses by suggesting those in your area

Comment

Inspiration

With the COVID-19 lockdown and economic crisis hurting business everywhere, we sought to create a chatbot that could help us connect you with local businesses that especially need support during this time. We've created and easy to use chatbot that will provide you with businesses for all your wants and needs, food, services, and more! To make a direct resource for people to find local businesses in their area to support, we developed Sully, a chatbot that gives you local business recommendations based on what you need. No more filtering through bigger businesses or chain companies to find local gems!

What it does

When the user enters their zip-code and the category of business they want to visit, Sully will return recommendations. More information is on our website

How we built it

Sully is a chatbot built with AI/ML through Dialogflow. After a user types something, it gets sent to a Flask script which processes the data, retrieves any necessary information from our Cloud Firestore database in Firebase, and returns it to Dialogflow. Our Flask script is deployed on Google Cloud App Engine and uses a webhook to send a fulfillment message to Dialogflow. To get the initial data in Firestore, we built a web scraper using python.

Challenges we ran into

This was our first time working with Dialogflow and we were not very proficient in Python so it was a challenge learning the technologies. To get Dialogflow, Flask, and Firebase connected, it was a lot of networking and sending data to the right locations and in the right format, which took time to figure out. Deploying Flask to Google App Engine was also difficult since we had little experience with GCS or Flask.

Accomplishments that we're proud of

We're proud that we were able to build a chatbot using AI for the first time and learn new technologies along the way.

What we learned

Michelle: I personally got more comfortable with Python, HTTP Requests, and learned Dialogflow. Since most of the technologies we used were apart of Google Cloud Services, I had to learn how to use the console and read the debug logs. I understand where data is actually going a bit better now.

Tracy: This was my first hackathon, so I learned so much, especially since this was my first time making a project that isn't a mobile app. I learned how to use Dialogflow and got more familiar with working with Cloud Firestore as well as the concept of web scraping.

We (the team) learned a lot about how important it is to choose the right technologies and outline our idea before building on it. Choosing the right languages and libraries to use and knowing where all the data is supposed to go makes coding a lot easier.

What's next for Sully

In the future, we want to expand Sully's database to more than the six cities we currently have. We also want to allow the user to have better conversations with her. Right now the only way to access Sully is through our website, but in the future we want to support more platforms.

Built With

Share this project:

Updates