Inspiration

All Brandon Wu seems to do is go to the gym, so we want to introduce him to a hobby diversi*fier* to help him branch out of his monotonous routine.

What it does

Hobbyfier takes in all of a user's Google Calendar events and analyzes their past activities using an LLM to recommend new hobbies based on their observed interests. For example, Brandon would sign into his Google account and receive recommendations to try yoga or hiking because he already loves fitness (a.k.a. his boring ex, the gym).

How we built it

After brainstorming our project details and outlining our web design using Figma, we attempted various web frameworks to add Google authentication to our web application for custom parsing through data from a user's Google Calendar. We combined an LLM with a database of activities to extract the user's Calendar events for their availability and past activities to predict what new things they would like to try. Finally, we constructed a front-end interface that dynamically updates based on each user's Google Calendar to provide more details on their recommended new hobbies with a supplemental calendar view.

Challenges we ran into

Initially, we had difficulty implementing Google authentication into our code. We were using React, which required too much redirecting in Express, so we transitioned to directly using HTML, CSS, and JavaScript instead. We also found it tricky to integrate the LLM with a database of recommended activities.

Accomplishments that we're proud of

We were able to figure out how to implement Google authentication and designate the process with its own redirected page. We are also proud of how we combined our dataset with an LLM to recommend relevant activities by extracting user data from Google Calendar. Last but not least, we love our UI design, including our typography, color palette, and different icon drawings of different hobbies!

What we learned

This was a huge learning experience for all of us, especially when it came to developing with an LLM while integrating Google Calendar. Throughout the process, we learned how to structure and execute a project on a team, adapting and evolving as different challenges arose.

What's next for Hobbyfier

We would love to incorporate weather information alongside our hobby recommendations to let users know when the best time to engage in the activity is. Additionally, we plan to develop a feature that enables users to combine their Google Calendar with a friend's to scan for overlapping interests to discover together during mutual availability.

Share this project:

Updates