HiveFive

Example Chats with Buzz Chatbot
GPS Information for Hive Members
Emotional Sentiment Analysis of Facial Response

\noindent \textbf{HiveFive} is a React app with Google Firebase (auth/real-time) and MongoDB that helps student friend groups form social groups called ``hives,'' chat, and quickly decide on plans. We were motivated by the struggles as students of event planning in our daily lives: meetups stall with busy schedules, indecision, bad timing, and distance. So, we built a tool that captures the right context at the right moment: personalized profiles (interests, majors, visibility controls, consent flags), an LLM chatbot that takes a meet request and asks follow-ups, facial-response visuals (specified in our design) to illustrate engagement while viewing an invite, a hive photo gallery to track participation, and a swiping flow that records quick yes/no decisions. Our app collects labeled data, tracking user reactions and geospatial data with respect to whether a user accepts or declines an event. We also incorporated head movement tracking and analysis collected by AirPods to assess reactions of users after events are suggested.

\noindent Our biggest challenges were product and ethics, not just code. Motivation-wise, we wanted planning to respect feasibility (free/busy, distance, price) and group opinions (interest vs.\ hesitation) while keeping users in control. Technical hurdles included combining Firebase real-time events with MongoDB schemas, keeping profile visibility rules correct, handling photo uploads and metadata safely, and making the LLM's questions feel helpful rather than spammy. Data hurdles included permission fatigue and noisy or missing sensor data. Through this process, we learned that our everyday devices, like phones and headphones, collect much more relevant data than we initially assume, and this data can be used in new ways to solve problems. We hope to use this information to build sma