RSF Now shows real-time occupancy at the RSF weight room and predicts how busy it will be for the next several months. It's a student side project with no university affiliation, no ads, and no monetization of any kind.
Current occupancy comes from the RSF's live sensor system. Predictions are generated by two machine learning models, a Random Forest and a neural network, trained monthly on over 177,000 historical readings. They account for the Berkeley academic calendar, so finals, dead week, breaks, and holidays are all factored in.
177k+
readings in training data
180
days predicted ahead
+/-9.4%
model error (MAE)
No sign-up, no tracking, no data collected about you. This will stay free.
Questions, bug reports, or ideas: lucaschang@berkeley.edu.