our logo!

Stryd

Wildhacks 2026 Project

Do you like waiting in line? Well, we don't either. While we can try to guess wait times for say TSA or the checkout line at Trader Joe's, it would help to be able to see what's actually going on. With this knowledge, we can help people plan out their days and their trips to minimize the amount of time spent waiting in line. Our frontend is built using React Native and is intented to be an iOS and Android app. For the backend, we used Supabase for authentication and to store the places the user frequents.

Here are some of the features of the homescreen.

stryd_demo_home

The algorithm we use to predict times is based on a quantile regression model. We would install cameras in the places of interest/connect to cameras in the places of interest to obtain a live feed. Then using this live feed, we would use a crowd density estimation model using deep learning to estimate the size of the crowd at each location. After recording a few weeks of data, we would use our quantile regression model to estimate the crowd size for each location at all times. A quantile regression model is better than a traditional linear regression model, because quantile regression is more robust to outliers. If some extraordinary event causes the foot traffic of a specific location to dramatically deviate from the amount predicted by the quantile regression model (ex. a free matcha popup at the mall), the app will update current wait times to match the data from the deep learning model, and the quantile regression model will not be as affected by this outlier, meaning future predictions will stay as accurate as possible.

Due to time/equipment/money restraints, we are not able to fully implement this plan, so we have included a small-scale demo of how the regression model and deep learning models work.