To follow in situ and in real time how catalyst surfaces respond to gas composition changes is a challenge. This study reports on an eventaveraging method, based on cyclic gas pulsing and software-based image recognition, that overcomes the challenge for large photoelectron spectroscopy datasets.