ABRLog: a multi-device android bugreport log dataset for intelligent analysis

Android devices generate a wealth of diagnostic data through their built-in logging subsystems, offering invaluable insights into application behavior, hardware performance, and operating system stability. However, the effective utilization of this data for intelligent analysis - particularly, for anomaly detection, failure diagnosis, and security monitoring - remains constrained by the scarcity of publicly available, well-structured, and multi-device log datasets. To address this gap, we introduce ABRLog (Android BugReport Log), a comprehensive multi-device log dataset derived from bug report captures collected across 15 distinct Android devices representing five major manufacturers (Samsung, Motorola, Google, Xiaomi, and Huawei) and spanning six Android versions, from 9 to 14. A total of 45 bugreport archives were systematically processed through two complementary extraction strategies: (i) line-by-line regular expression matching for granular event extraction; and (ii) section-based delimiter parsing for high-level block segmentation. This dual approach yielded 1012 structured log files, organized into 24 semantically coherent block types. The final dataset comprises approximately 10 million log lines (1.14 GiB), extracted from 6.34 GiB of raw bugreport data. By providing a systematically curated, multi-vendor, multi-version log corpus, ABRLog aims to serve as a foundational resource for the development and benchmarking of machine learning models for log-based anomaly detection, predictive maintenance, and intelligent system diagnostics on Android platforms.