We collected usage information from 12,500 Android devices in the wild over the course of nearly 2 years. Our dataset contains 53 billion data points from 894 models of devices running 687 versions of Android. Processing the collected data presents a number of challenges ranging from scalability to consistency and privacy considerations. We present our system architecture for collection and analysis of this highly-distributed dataset, discuss how our system can reliably collect time-series data in the presence of unreliable timing information, and discuss issues and lessons learned that we believe apply to many other big data collection projects.
No takes yet. Share an insight, caveat, or question.
Wagner et al. (2014) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: