You are here

» Julie Kientz Research Group

Julie Kientz Research Group

Spring 2014

Can Smartphone Usage Predict Sleep Status?

Sleep is an important component of health. With long-term, behavioral sleep issues such as insomnia, monitoring the amount of sleep you get can be an important part of helping to find the causes of the problems and work on a solution. Current sleep sensors often require on-the-body sensors and that the user must remember to put it on or turn it on every night, which can lead to high user burden and unreliable data. These issues make it difficult for users to learn about their sleep behaviors over the long term.

In this directed research group, we will be working on a new way of identifying sleep behaviors by looking at how people use their smart phones. For example, many people have the habit of charging their phone while they sleep, or the phone may not move for several consecutive hours. The work involved will be collecting data via phone usage logs as well as “ground truth” data via commercial sleep sensors and manual sleep diaries, and then applying machine learning techniques (using the Weka toolkit) to determine whether we can build a model that predicts a user’s sleep status. We already have access to an Android-based logging tool and commercial sleep sensors, so part of the work for this group will be to recruit participants, set up the logging tool on their phone, give them the sleep sensors, and collect information from them and analyze it.

Students participating would benefit from having some familiarity with Android phones, basic programming skills (in case we need to do any modifications to the logging tool), and quantitative data analysis skills (or a willingness to learn).

Meeting times for Spring 2014 are Thursdays from 3:30-5:00. The group is currently full for Spring 2014.


Julie Kientz's Directed Research Group archive: