Personal big data analytics unveiled usage behaviors related to smartphone addiction

KAIST researchers unveiled unique usage behaviors related to smartphone addiction by utilizing personal big data analytics.

Article  |  Fall 2014

As the popularity of smartphones has been rapidly increasing in recent years, negative aspects of their usage have emerged, such as social conflicts, sleep deprivation, and attention deficits. Prof. Uichin Lee and his colleagues at KAIST identified usage behaviors related to smartphone addiction. The results were published at the ACM SIGCHI conference, which is a premiere conference on Human Computer Interaction (HCI) [1]. In the past, psychologists mostly used survey questionnaires to understand usage behaviors related to addiction, which are subjective and prone to recall errors. In contrast, Prof. Lee’s team aimed at overcoming such limitations by analyzing personal big data of smartphone usage.

Prof. Lee’s team collected actual smartphone usage data from 95 college students (over 50,000 hours of usage data) in 2012. The study participants were divided into risk and non-risk groups based on the scores of a smartphone addiction questionnaire by Korea’s National Information Society Agency (NIA). The team then performed exploratory statistical analyses to uncover usage behaviors related to smartphone addiction.

The team showed that the risk group’s smartphone usage duration per day was longer than the non-risk group’s (risk group: 4 h 13 min vs. non-risk group: 3 h 27 min), and significant usage time differences were observed in the morning and evening hours. Despite the fact that smartphones offer various functions, the risk group spent significantly more time on a few apps than the non-risk group (e.g., KakaoTalk, Facebook). App notifications (e.g. message arrivals in KakaoTalk) were closely related to addictive usage behavior; the risk group spent 38 minutes longer in the usage sessions that began with notified apps. That is, for those who lack self-regulation, notifications may serve as external stimulus that leads to more frequent usage of smartphones.

Moreover, Prof. Lee’s team proposed a novel method that can automatically identify whether a user belongs to the risk or non-risk group at an accuracy level of 85% by using personal big data. This method will pave the way for early detection of smartphone addiction, which will help those who are at the early stages of addiction to receive timely care services.

The team is currently developing usage intervention apps for smartphone overuse. The team recently reported the survey results of various intervention apps available on Google Play and the Apple App Store [2]. In the smartphone non-use workshop at the CHI 2014 conference, the team identified how people typically coped with interferences due to smartphone overuse and showed that designing systematic tools to support temporary non-use of smartphones would be very important for usage intervention [3].

[1] U. Lee, J. Lee, M. Ko, Ch. Lee, Y. Kim, S. Yang, K. Yatani, G. Gweon, K.-M. Chung, and J. Song, “Hooked on Smartphones: An Exploratory Study on Smartphone Overuse among College Students,” ACM SIGCHI Conference on Human Factors in Computing Systems (CHI’14), Toronto, Canada, April 26-May 1, 2014.

[2] M. Ko, J. Lee, S. Yang, and U. Lee, “An Analysis of Mobile Apps for Intervening Excessive Smartphone Usage: Intervention Method Perspectives,” HCI Korea, Feb. 2014.

[3] U. Lee, S. Yang, M. Ko, and J. Lee, “Supporting Temporary Non-Use of Smartphones,” Refusing, Limiting, Departing: Why We Should Study Technology Non-use. (In conjunction with CHI 2014), Toronto, Canada, April 26, 2014.