AffectLogger audience response dataset: subjective event timing recordings from a live comedy performance

This dataset contains event-level audience response data recorded using the AffectLogger system during a live comedy performance on October 16, 2025. AffectLogger is a smartphone-based system designed to record the timing of subjective experiences in naturalistic settings with minimal participant burden. During the approximately 105-minute performance, ten audience members used a simple on/off interface implemented on their smartphones to indicate periods in which they recognized themselves as laughing. The resulting data consist of timestamped response states recorded at the device logging rate (approximately 25–30 Hz). Each CSV file corresponds to one participant and contains raw event logs rather than uniformly sampled time series. Because the data are recorded as raw device logs, multiple records may occasionally share identical timestamps. Such duplicated timestamps reflect the logging process rather than distinct events and can be removed or merged during preprocessing. In the associated study, laughter events are defined as transitions from false to true in the response state, allowing event-level analyses of audience reactions. The dataset supports analyses of temporal response patterns such as inter-event intervals (IEI), local variability (LV), and other event-sequence–based methods for studying collective audience dynamics. This dataset accompanies the manuscript: Nomura, R., & Yamazato, T.AffectLogger: A smartphone system for recording subjective event timing. To protect participant privacy, video recordings used for facial expression analysis are not included in this dataset.