Characterizing Smartphone Usage Patterns from Millions of Android Users

Abstract

The prevalence of smart mobile devices has promoted the popularity of mobile applications (a.k.a. apps). Supporting mobility has become a promising trend in software engineering research. This article presents an empirical study of behavioral service profiles collected from millions of users whose devices are deployed with Wandoujia, a leading Android app-store service in China. The dataset of Wandoujia service profiles consists of two kinds of user behavioral data from using 0.28 million free Android apps, including (1) app management activities (i.e., downloading, updating, and uninstalling apps) from over 17 million unique users and (2) app network usage from over 6 million unique users. We explore multiple aspects of such behavioral data and present patterns of app usage. Based on the findings as well as derived knowledge, we also suggest some new open opportunities and challenges that can be explored by the research community, including app development, deployment, delivery, revenue, etc.Index Terms-mobile apps, app store, user behavior analysis !

Urban Vibes and Rural Charms: Analysis of Geographic Diversity in Mobile Service Usage at National Scale

“…Geographical diversity has been often overlooked also in prior explorations of mobile data traffic considering multiple services. Most works in the literature have a different focus, including differences in mobile application usage in time [35] or across the subscriber population [21]. Attention has also been paid to patterns in the utilization of apps by individual users, finding, e.g., that mobile service usage is very heterogeneous among the user base [13], strongly depends on context [6], is characterized by brief bursts of interactions [14], and is influenced by the type of device used [19].…”

Deriving User Preferences of Mobile Apps from Their Management Activities

“…Essentially, the data used in our study are also collected through the log of Wandoujia's management app. The app usage log collected through Wandoujia is by far the largest in the literature, covering millions of users and millions of apps (a sample of the data has been released for public research [Li et al 2015b]). Many interesting research problems can be pursued based on such a dataset.…”

“…The data are legally used without leaking any sensitive information. A sample of the dataset has been released, along with our previous work [Li et al 2015b], which can be accessed at http://www.sei.pku.edu.cn/∼ liuxzh/appdata. For more information, please contact liuxuanzhe@pku.edu.cn.…”

On the feasibility of a user-operated mobile content distribution network

“…We approximate application popularity by a Pareto distribution, given ample evidence from related literature that smartphone application popularity follows such distribution, e.g., [13], [14]. According to the used Pareto distribution, 28% of the total number of installed applications in the mobile devices of the destination nodes are instances of the most popular application, whereas the least popular application counts only 4.5% of that number (i.e., skewed, heavy-tailed distribution, where some applications are installed in almost all devices, whereas the remaining are present in a smaller number of devices).…”

“…The increased demand for mobile data and the corresponding relatively high skewness in the popularity of the requested content [13] [14] has led to an increased interest in the area of data offloading through opportunistic communication with or without the assistance of the cellular infrastructure. For example in [27] authors present a cellular assisted mechanism to serve user requests from other mobile users located geographically close, by clustering crowded places in dataspots and by tracking the location of users, as well as the content cached in their devices.…”