An open-source framework for context-aware monitoring of mobile application feature usage

Abstract

Mobile devices such as smartphones and tablets are widely used globally. There are diverse segments of mobile users and a huge number of mobile applications available in the various mobile app stores. Moreover, the number of smart phone users and number of mobile applications in mobile app stores are growing rapidly. In such a situation, it is challenging for mobile application to keep up the interest of mobile users alive. In fact, study shows that the majority of the mobile applications are having a short life span.

For mobile application to survive, experts emphasize on understanding the mobile users' interests and their usage pattern. With better insight of the usage patterns, application designers can align the future application development as per the target user groups and their interests.

In this thesis work, design science methodology is used for creating an artefact, monitoring framework to monitor the application feature usage along with the user context information for instance device acceleration, device orientation, network connectivity (such as 3G, Wi-Fi), attached peripheral devices. Literature study is conducted to understand the problem domain, user context elements, capabilities of Android device sensors, and Android framework. A systematic architecture design process is used to develop the architecture of monitoring framework. The monitoring framework is an open-source solution for Android devices and utilizes another open-source framework, Funf framework to collect the information from various built-in Android device sensors.

The developed software artefact is evaluated through a requirements' checklist based architectural evaluation technique, a performance test of developed artefact, and a real field study. The results of field study demonstrate that monitoring framework helps in collecting the granular details of application feature usage along with users' context information. Moreover, performance testing results illustrate that the data collection through monitoring framework adds a very low latency (around 2-3 milliseconds) on the response times of the Android application's monitored features.

This design science research in this thesis has importance from two aspects. First, it has a high practical relevance. The developed artefact can measure application feature usage along with users' context information. This information is useful in identifying usage patterns which in turn help in planning for future application development and marketing of application. Second, this research can be beneficial to the theory building in the emerging research field of mobile application analytics. The developed artefact, branded as ContextLogger3, is currently utilized successfully for a few research projects and studies at Aalto University.