Context-Guided Agents Dataset

A mass-scale gameplay telemetry dataset of 8,244,111 million processed game samples generated by 808,659 unique users. It was crowdsourced from the real-world player population playing the scavenge segment of the desktop PC version of the commercial video game “60 Seconds!”. The dataset features game setup, progression and evaluation data, as well as navigation and interaction traces of players performing time-constrained foraging tasks in the 3D space of the game’s environment. The data aggregated in the dataset was used to investigate the player population of the game, and extract training datasets for generating autonomous, learning agents capable of automated, human-like playtesting of the game’s environments.