Ahead of the Sweep, Not Ahead of the Mess

Forecasting Hot-Desk Tidiness-Sweep Failures from Swipe-In Time Regularity Rather Than Logged Hours

A paper from the School of Emergent Priorities reports a fourteen-week deployment of a Swipe Regularity Index, adapted from circadian rest-activity-rhythm methodology, forecasting hot-desk tidiness-sweep failure across 620 desk-weeks on two hot-desk floors. A regularity-only model reaches an AUC of 0.71 against 0.54 for Facilities' existing hours-logged risk score, and a feature ablation finds logged hours adds no detectable forecasting value once regularity is available.

@misc{slop_h4dvr2,
  author       = {Fenna Okoro and Kwame Lindqvist},
  title        = {Ahead of the Sweep, Not Ahead of the Mess: Forecasting Hot-Desk Tidiness-Sweep Failures from Swipe-In Time Regularity Rather Than Logged Hours},
  year         = {2026},
  publisher    = {Slop University},
  doi          = {10.5555/slop.h4dvr2},
  url          = {https://slop.university/outputs/slop-paper-forecasting-which-hot-h4dvr2/},
  version      = {1.0},
  note         = {Research paper},
}