Faces oF soFTware analyTics

The sofTware developmenT process produces various types of data such as source code, bug reports, checkin histories, and test cases. Over the past few decades, researchers have used analytics techniques on such data to better understand software quality (for some examples, see the sidebar). Also during that period, Internet access has become widely available, and software services now play an indispensable role in everyday life. Much richer types of data at much larger scales now help practitioners better understand how software and services perform in real-world settings and how end users employ them.1 To address the increasing importance and abundance of data in the software domain, we formed Microsoft Research Asia’s Software Analytics Group (http://research.microsoft. com/groups/sa) in 2009. We proposed the concept of software analytics to expand the scope of analyzing software artifacts. Software analytics aims to obtain insightful and actionable information from software artifacts that help practitioners accomplish tasks related to software development, systems, and users. Although much research has covered software analytics, little work has covered its impact on software development practice.2,3 Software analytics can affect software development practice in different ways.4–6 For example, • software practitioners could conduct software analytics by themselves, • researchers from academia or industry could collaborate with software practitioners from software companies or open source communities and transfer their analytics technologies to real-world use, or • practitioners could organically adopt analytics technologies developed by researchers.