Weclome to UbiComp/ISWC 2019 Tutorial on Smartphone App Usage, Understanding, Modelling, And Prediction

See all other UbiComp/ISWC 2019 Tutorials.


Based on our past research in this area, we would like to run the first full day tutorial of Smartphone Apps Usage Understanding, Modelling and Prediction, which targets the audience of both researcher, engineers, and graduate student. In a nutshell, this tutorial aims to provide an in-depth and solid introduction on applying data science and data mining to smartphone app suage modeling through comprehensive theory and technical details as well as through detailed examples. In order to attract more participants and researchers to this tutorial, we will share two app datasets in terms of both short-term context-aware data and long-term usage dataset. This will provide a foundation for potential researchers interested in this area and a forum for the participants to communicate and discuss issues to promote the emerging research field.


The wide adoption of mobile devices and smartphone applications (apps) has enabled highly convenient and ubiquitous access to Internet services. For both app developers and service providers, it becomes increasingly important to predict how users use mobile apps under various contexts. Mining and learning from smartphone apps for users is an important and emerging research field. Research on mobile apps is beginning to grow up fast in the recent years. During the past four years of UbiComp, there are more than 10 papers on analysis and mining of smartphone apps, even one of them won the Best Paper Award of UbiComp2016.

Mining and learning from smartphone apps for users is very relevant to ubiquitous computing. Smartphone app are ubiquitous in our daily life. Abundant apps provide useful services in almost every aspect of modern life. Easy to download and often free, apps can be fun and convenient for playing games, getting turn-by-turn directions, and accessing news, books, weather, and more. Apps on smartphones can be considered as the entry point to access everyday life services such as communication, shopping, navigation, and entertainment. Since a smartphone is linked to an individual user, apps in smartphones can sense users’ behavior and activities. Researchers use the data recorded by smartphone apps to analyze apps and understand users. However, there are still not yet so many researchers in this emerging field. One reason is the dataset available for research and the second reason is the area is relatively new. As the researchers working in this area with more than 5 years and publishing more than 30 related high quality papers in this area in the top conference, we would like to provide a tutorial will provide a compact platform to help researchers to focus on this research area, and also open several dataset in this tutorial.

This tutorial is helpful for researchers to learn the basic idea and techniques in this area, and also distribute the latest progress on this hop topic, to promote the research area. First, the participants of this workshop can learn the data, methods, tools, and experiences from the analysis of smartphone apps. Second, the findings and discussions that are presented in the tutorial can motivate researchers. Third, by bringing together participants with a variety of backgrounds and goals, this tutorial provides a platform for interdisciplinary cooperation and networking.

Intended Audience and Required Knowledge

This tutorial is in the area of context-aware smartphone app usage behavior modeling and prediction. which is a popular and also relative new topic in the area of ubiquitous computing. The prerequisite knowledge for the audience includes the understanding of the basics of mobile system and networking, data science and data mining, and basic technical foundations of machine learning.

Brief Biography of Presenters

Sasu Tarkoma

Sasu Tarkoma is a Professor and Head the Department of Computer Science at the University of Helsinki. His research interests are data science, Internet technology, distributed systems, data analytics, and mobile and ubiquitous computing. His research has received several Best Paper awards and mentions, for example at IEEE PerCom, ACM CCR, and ACM OSR. His major achievements include the Carat energy profiler project with over 850 000 users globally. The Carat project has received the prestigious Mark Weiser Best Paper award at IEEE Percom 2015.

Vassilis Kostakos

Vassilis Kostakos is a Professor in Human-Computer Interaction at the University of Melbourne School of Computing and Information Systems. He is a Marie Curie Fellow, a Fellow in the Academy of Finland Distinguished Professor Program, and a Founding Editor of the PACM IMWUT journal. He holds a PhD in Computer Science from the University of Bath. His research interests include ubiquitous computing (Ubicomp), human-computer interaction (HCI), social computing, and Internet of Things.

Yong Li

Yong Li is currently an Associate Professor of the Department of Electronic Engineering, Tsinghua University. He received the B.S. degree in electronics and information engineering from Huazhong University of Science and Technology, Wuhan, China, in 2007 and the Ph.D. degree in electronic engineering from Tsinghua University, Beijing, China, in 2012. Dr. Li is one of Associate Editors of IMWUT. His papers have total citations more than 6400. Among them, ten are ESI Highly Cited Papers in Computer Science, and four receive conference Best Paper (run-up) Awards.

Sha Zhao

Sha Zhao is currently a Postdoctoral Research Fellow of the College of Computer Science and Technology, Zhejiang University. She received the Ph.D. degree from Zhejiang University, Hangzhou, China, in 2017. Her research interests include ubiquitous computing, mobile sensing, data mining, and machine learning. She focuses on mining and learning from smartphone apps, and one of her work won the Best Paper Award of UbiComp’16.

Format and Schedule

This tutorial will be given by five talks in about 4 hours, and each talk will be given by one speakers. The detailed scheduling is as follows.

Dataset and Publications



The online version of the slides is available at this link.