Conference PaperPDF Available

How Age and Gender Affect Smartphone Usage

  • September 2016
DOI:10.1145/2968219.2971451
  • Conference: UbiComp '16
Authors:
Ionut Andone at University of Bonn
Ionut Andone
Konrad Błaszkiewicz at University of Bonn
Konrad Błaszkiewicz
Mark Eibes at University of Bonn
Mark Eibes
Boris Trendafilov at University of Bonn
Boris Trendafilov
Show all 6 authorsHide
Download full-text PDF
Read full-text
Download citation

Abstract and Figures

Smartphone usage is a hot topic in pervasive computing due to their popularity and personal aspect. We present our initial results from analyzing how individual differences, such as gender and age, affect smartphone usage. The dataset comes from a large scale longitudinal study, the Menthal project. We select a sample of 30, 677 participants, from which 16, 147 are males and 14, 523 are females, with a median age of 21 years. These have been tracked for at least 28 days and they have submitted their demographic data through a questionnaire. The ongoing experiment has been started in January 2014 and we have used our own mobile data collection and analysis framework. Females use smartphones for longer periods than males, with a daily mean of 166.78 minutes vs. 154.26 minutes. Younger participants use their phones longer and usage is directed towards entertainment and social interactions through specialized apps. Older participants use it less and mainly for getting information or using it as a classic phone.
Figures - uploaded by Ionut Andone
Author content
All figure content in this area was uploaded by Ionut Andone
Content may be subject to copyright.
Content uploaded by Ionut Andone
Author content
All content in this area was uploaded by Ionut Andone on Nov 08, 2017
Content may be subject to copyright.
How Age and Gender Affect
Smartphone Usage
Ionut Andone
Konrad Błaszkiewicz
Mark Eibes
Boris Trendafilov
Institute of Computer Science
University of Bonn
Bonn, Germany
andone@cs.uni-bonn.de
blaszkie@cs.uni-bonn.de
mark.eibes@gmail.com
trendafilov.boris@gmail.com
Alexander Markowetz
markowetz.de
Bonn, Germany
alexander@markowetz.de
Christian Montag
Institute of Psychology and
Education
Ulm University
Ulm, Germany
christian.montag@uni-ulm.de
Permission to make digital or hard copies of part or all of this work for personal or
classroom use is granted without fee provided that copies are not made or distributed
for profit or commercial advantage and that copies bear this notice and the full citation
on the first page. Copyrights for third-party components of this work must be honored.
For all other uses, contact the Owner/Author.
Copyright is held by the owner/author(s).
UbiComp/ISWC ’16 Adjunct , September 12-16, 2016, Heidelberg, Germany
ACM 978-1-4503-4462-3/16/09.
http://dx.doi.org/10.1145/2968219.2971451
Abstract
Smartphone usage is a hot topic in pervasive computing
due to their popularity and personal aspect. We present
our initial results from analyzing how individual differences,
such as gender and age, affect smartphone usage. The
dataset comes from a large scale longitudinal study, the
Menthal project. We select a sample of 30,677 participants,
from which 16,147 are males and 14,523 are females, with
a median age of 21 years. These have been tracked for at
least 28 days and they have submitted their demographic
data through a questionnaire. The ongoing experiment has
been started in January 2014 and we have used our own
mobile data collection and analysis framework. Females
use smartphones for longer periods than males, with a daily
mean of 166.78 minutes vs. 154.26 minutes. Younger par-
ticipants use their phones longer and usage is directed to-
wards entertainment and social interactions through spe-
cialized apps. Older participants use it less and mainly for
getting information or using it as a classic phone.
Author Keywords
Mobile devices; smartphone usage; mobile computing; user
behavior observation; data mining.
ACM Classification Keywords
H.5.m [Information interfaces and presentation (e.g., HCI)]:
Miscellaneous
10.72. Tables 1 and 2 present the age and gender distribu-
tion of our dataset, alongside data from the World Factbook
[5]. We use the data available for Germany since the major-
ity of our users come from this country. We can see that our
dataset is skewed towards the younger population. Some
parents have installed our app on the phones of their chil-
dren, but this does not constitute the majority. The app has
been used as a method of "time on device" management.
Since the gender and age was collected through our in-app
questionnaire, we do not have a guarantee that all of our
users provided correct data. Nonetheless, we have a large
enough sample for this to not be statistically significant.
0
20
40
60
female male
Gender
Time (minutes per day)
Categories
Communication
System
Social
Unknown
Games
Media.and.Video
Other
Figure 2: Daily mean phone usage duration broken down into
different app categories for males and females. Females use
communication and social apps longer than males.
0
500
1000
Time (minutes per day)
Age Groups
0−11y
12−17y
18−25y
26−35y
36−50y
51y+
Figure 3: For how long different
age groups use the phone daily.
Teenagers (12-17 years) lead in
usage, with around 190 minutes.
Usage decreases with increase of
age.
The app categories have been modeled following the main
categories from the Google Play Store. The communication
category contains messaging apps as well as web browsers
and email clients ( e.g. WhatsApp, Facebook Messenger,
Chrome, Gmail). The social category contains social net-
works such as Facebook, Snapchat and Instagram. The
system category includes apps which are not available on
the Google Play Store because they are part of the Android
OS or vendor specific extension (e.g. different system set-
tings, SMS, Phone apps, etc.). The most important apps in
this category are launchers and lockscreen apps. An app
was placed in the unknown category when this app was not
available from the play store and was not placed in the rest
of the categories. An app was placed in the other category
when its category was known but was not one of the main
categories we used in the plots. All of the other categories
are self explanatory.
0
20
40
60
Communication
System
Social
Unknown
Games
Media.and.Video
Other
Category
Time (minutes per day)
Age groups
0−11y
12−17y
18−25y
26−35y
36−50y
51y+
Figure 4: The influence of age on daily average phone usage
clustered by different app categories. Communication and social
apps are used heavily by teenagers. Games, Media and Video
apps usage decline as age increases.
Phone usage differs based on the gender of the participant.
Females spend more time on their phones than males, they
spend on average per day 166.87 minutes (SD = 91.95),
while males spend 154.26 minutes (SD = 92.78). On aver-
age, women spend more time in communication and social
apps while men spend more time playing games . Different
age groups show even bigger differences. The daily mean
phone usage is highest for teenagers, 12 to 17 years old
with 193.64 minutes, and goes down while age increases
to 117.95 minutes at 51+ years old. A similar pattern can
be observed for the usage of communication and social
apps, where in both of these cases, teenagers are the most
prominent users. The situation looks different for gam-
ing apps. Users in our youngest groups spend the same
amount of time in both gaming and using communication
apps. On average, this sums up to over 40 minutes daily,
while other age groups spend around 20 minutes daily with
games. Young users, under 18 years, spend much more
time in media and video apps compared to older ones.
Participants over 30 years old spend less than 10 minutes
watching videos and media.
Conclusions
We clearly see that age and gender play a big role in the
amount and type of phone and app usage. Younger peo-
ple are communicating more and thus are using social
and communication apps more often. They also play more
games, especially children, than the rest of the population.
Older people have different needs and often use the smart-
phone just as a method of informing themselves. Either
by reading the news or communicating with their peers by
using the smartphone as a regular phone. As future work,
we want to expand the current analysis to include other
events such as phone locks/unlocks, movement patterns,
app installs, and activity patterns. Besides age and gender
we plan to look into how education and personality affects
phone usage. We hope that our analysis results might influ-
ence future designs for operating system interfaces, based
on differences between individuals.
REFERENCES
1. Ionut Andone, Konrad Błaszkiewicz, Mark Eibes, Boris
Trendafilov, Christian Montag, and Alexander
Markowetz. 2016a. Menthal - Running a Science
Project as a Start-Up. In Computing in Mental Health,
Workshop at CHI 2016. ACM.
2. Ionut Andone, Konrad Błaszkiewicz, Mark Eibes, Boris
Trendafilov, Christian Montag, and Alexander
Markowetz. 2016b. Menthal: A Framework for Mobile
Data Collection and Analysis. In Ubicomp/ISWC 2016
Adjunct. ACM.
3. Ricardo Baeza-Yates, Di Jiang, Fabrizio Silvestri, and
Beverly Harrison. 2015. Predicting the next app that
you are going to use. In Proc. WSDM 2015. ACM,
285–294.
4. Matthias Böhmer, Brent Hecht, Johannes Schöning,
Antonio Krüger, and Gernot Bauer. 2011. Falling
asleep with Angry Birds, Facebook and Kindle: a large
scale study on mobile application usage. In Proc.
MobileHCI 2011. ACM, 47–56.
5. CIA. 2015. The World Factbook 2014-15. GPO.
6. Ericsson. 2016. Ericsson Mobility Report. (2016).
7. Denzil Ferreira, Anind K Dey, and Vassilis Kostakos.
2011. Understanding human-smartphone concerns: a
study of battery life. In Pervasive computing. Springer,
19–33.
8. Andrea Girardello and Florian Michahelles. 2010.
AppAware: which mobile applications are hot?. In Proc.
MobileHCI 2010. ACM, 431–434.
9. Daniel T Wagner, Andrew Rice, and Alastair R
Beresford. 2014. Device Analyzer: Large-scale mobile
data collection. ACM SIGMETRICS Performance
Evaluation Review 41, 4 (2014), 53–56.
... It demonstrated that age and gender affected smartphone usage, and also contributed to the significance in results. It is supported by research which reported that younger adults use their phones for longer periods of time, and their usage is primarily directed toward entertainment and social interactions via specialised apps [70]. Older participants use them less frequently, and primarily for information or as a traditional phone [70]. ...
... It is supported by research which reported that younger adults use their phones for longer periods of time, and their usage is primarily directed toward entertainment and social interactions via specialised apps [70]. Older participants use them less frequently, and primarily for information or as a traditional phone [70]. Furthermore, females use smartphones for longer periods than males, with an average of 166.78 min vs. 154.26 ...
... Furthermore, females use smartphones for longer periods than males, with an average of 166.78 min vs. 154.26 min per day, respectively [70]. In terms of purchasing behaviours, female purchasers opted to choose healthier food options such as vegetables and fruits, when compared to male purchasers [71]. ...
Effectiveness of Integrated Technology Apps for Supporting Healthy Food Purchasing and Consumption: A Systematic Review
Article
Full-text available
  • Aug 2021
A healthy diet is essential for good health and nutrition, though literature showed that there are various factors affecting the intention to purchase and consume healthy food. Technology integration is known to be useful in various aspects, but findings from studies on the efficacy of technology integration to improve healthy food consumption and purchase have largely been inconsistent. Therefore, we aimed to examine the efficacy of interventions that use technology apps to improve healthy food purchasing and consumption in adults. Relevant studies were identified through PubMed, Scopus, CINAHL, SportDiscuss and ACM Digital Library. Twenty studies were included in the systematic review. The majority of studies (n = 18) used a smartphone in the intervention, and only two studies used a personal digital assistant. The results showed that technology integration-based intervention favoured healthy changes in household food purchases, and increased consumption of healthy food and healthy eating outcomes – albeit to different extents. Overall, technology apps are convenient and user-friendly tools to encourage a change in healthy food purchase and consumption among people.
... Given that the smartphone is a multipurpose device, which can be used in business/school/university and personal contexts, we aimed to investigate cross-sectional predictive effects of the aforementioned TAM variables on smartphone use (and disorder tendencies) in both contexts. Given the recent inconclusive results of studies on gender differences in technology acceptance and use of technologies such as the smartphone (Andone et al., 2016;Lachmann et al., 2017;Peterka-Bonetta et al., 2019;Sindermann et al., 2020), as well as those examining associations for age with technology and smartphone use (Lachmann et al., 2017;Peterka-Bonetta et al., 2019), age and gender were additionally included in the models (see Figure 1). ...
Linking the Technology Acceptance Model to Smartphone Use and Smartphone Use Disorder: Results from a Survey Study
Article
  • Apr 2021
Background: Despite the benefits that may result from smartphone use, evidence increasingly indicates that smartphone use may also have negative consequences when used in a disordered manner. One major concept in this research domain is a putative smartphone use disorder. Objective: It is not known how a positive evaluation or acceptance of the smartphone, indicated for example by Technology Acceptance Model (TAM) variables, is associated with its increasing use and consequently tendencies towards smartphone use disorder. Methods: To close this research gap, an online survey with N=698 smartphone users (n=330 men, n=368 women) was conducted to study potential links. All participants completed a TAM measure, provided information on daily smartphone use for personal and business use, and completed a scale assessing tendencies towards smartphone use disorder. Results: Overall, TAM variables were positively related to smartphone use and tendencies towards smartphone use disorder. Descriptively stronger correlations of TAM with smartphone use were found in the personal use context compared to the business use context. Moreover, significantly stronger correlations of TAM scales with smartphone use disorder tendencies were found compared to the associations between TAM scales and daily smartphone use. Moreover, user gender also played a role in these relationships. Conclusion: Potential explanations for the exploratory findings are discussed and limitations and potential avenues for future research are provided, such as conducting longitudinal studies to causally investigate the associations between TAM and smartphone use (disorder tendencies).
... In total, we conducted seven virtual workshops à four participants using the software Zoom. Due to the use of apps by younger people [57], we set the focus on this group and recruited the participants via the online platforms of ETH Zurich, the University of Zurich, and the Swiss Study Foundation. Interested persons then received an information sheet and a consent form. ...
Why should I use a multi-hazard app? Assessing the public’s information needs and app feature preferences in a participatory process
Article
Full-text available
  • Mar 2021
In the age of nearly instant information and high-bandwidth communication, public expectations regarding the availability of authoritative short-term and real-time hazard information are evolving rapidly. Despite the existence of numerous apps covering hazard information, little research has been conducted to assess what the public actually expects from such an app. Perceived usefulness is crucial as the public will only use apps offering an added value to them. Several studies have identified that people prefer an app combining relevant hazards instead of single hazard apps. Therefore, we focused our analysis on multi-hazard apps. With seven virtual interactive workshops, we assessed the public’s information needs and app feature preferences. With respect to the content of multi-hazard apps, our results show that participants would mainly combine natural hazards but also have an interest in integrating anthropogenic and socio-natural hazards. To this end, participants advocate sending push notifications for low-probability hazards via general-purpose apps (e.g., weather apps) to reach a larger audience. Despite the desired diversity of hazard information, participants wish for an app focusing on relevant information only and redirecting users to other authorized sources for further information. Participants define the following as relevant information: location, time, hazard severity, behavioral recommendations and the contact details of emergency services. In addition, they wish for the following features: push notifications, a help button, a sharing function, a chat forum, an “I am safe” button, and a “Report an event” button.
Diagnostic Accuracy of Smartphone-Based Audiometry for Hearing Loss Detection: Meta-analysis
Article
  • Sep 2021
Background Hearing loss is one of the most common disabilities worldwide and affects both individual and public health. Pure tone audiometry (PTA) is the gold standard for hearing assessment, but it is often not available in many settings, given its high cost and demand for human resources. Smartphone-based audiometry may be equally effective and can improve access to adequate hearing evaluations. Objective The aim of this systematic review is to synthesize the current evidence of the role of smartphone-based audiometry in hearing assessments and further explore the factors that influence its diagnostic accuracy. Methods Five databases—PubMed, Embase, Cochrane Library, Web of Science, and Scopus—were queried to identify original studies that examined the diagnostic accuracy of hearing loss measurement using smartphone-based devices with conventional PTA as a reference test. A bivariate random-effects meta-analysis was performed to estimate the pooled sensitivity and specificity. The factors associated with diagnostic accuracy were identified using a bivariate meta-regression model. Study quality was assessed using the Quality Assessment of Diagnostic Accuracy Studies-2 tool. Results In all, 25 studies with a total of 4470 patients were included in the meta-analysis. The overall sensitivity, specificity, and area under the receiver operating characteristic curve for smartphone-based audiometry were 89% (95% CI 83%-93%), 93% (95% CI 87%-97%), and 0.96 (95% CI 0.93-0.97), respectively; the corresponding values for the smartphone-based speech recognition test were 91% (95% CI 86%-94%), 88% (95% CI 75%-94%), and 0.93 (95% CI 0.90-0.95), respectively. Meta-regression analysis revealed that patient age, equipment used, and the presence of soundproof booths were significantly related to diagnostic accuracy. Conclusions We have presented comprehensive evidence regarding the effectiveness of smartphone-based tests in diagnosing hearing loss. Smartphone-based audiometry may serve as an accurate and accessible approach to hearing evaluations, especially in settings where conventional PTA is unavailable.
A mindfulness meditation mobile app improves depression and anxiety in adults with sleep disturbance: Analysis from a randomized controlled trial
Article
Objective The objective of this study was to 1) determine the effects of a meditation app on depression and anxiety in adults with sleep disturbance, and 2) explore the potential mediating effects of fatigue, daytime sleepiness, and pre-sleep arousal on the relationship between use of the meditation app and changes in depression and anxiety. Methods Participants were 239 adults with elevated insomnia symptoms (i.e., scores ≥ 10 on the Insomnia Severity Index) and limited or no previous experience with meditation. Depression, anxiety, fatigue, daytime sleepiness, and pre-sleep arousal were assessed at baseline, four weeks, and eight weeks. Repeated-measures ANCOVAs assessed intervention effects on depression and anxiety. Mediation models were estimated using the PROCESS macro. Results Participants in the meditation group had more improvement in depression and anxiety symptoms during the intervention period than did those in the control group. Changes in somatic and cognitive pre-sleep arousal at mid-intervention fully mediated effects on depression and partially mediated effects on anxiety. There were no significant indirect effects of fatigue and daytime-sleepiness on changes in mental health. Conclusions A meditation app may improve depression and anxiety in adults with sleep disturbance, with effects being driven by improvements in pre-sleep arousal. Future studies should consider targeting pre-sleep arousal to improve mental health in this population.
PREDICTING STUDENTS' MULTIPLE INTELLIGENCE FROM BEHAVIORAL PATTERNS OF SMARTPHONE USAGE
Preprint
Full-text available
  • Aug 2021
Over the years, understanding students' intelligence has become a crucial issue for educators. With available theories on multiple intelligence as opposed to unitary cognitive ability, educators are confronted with more challenging task of measuring students' intelligence and their overall information processing capabilities. Due to ubiquitous nature of smartphone technology which has led to its wider use among digital natives, researchers have begun to statistically predict students' academic achievement and learning behavior using streams of smartphone sensing data. However, predicting their intelligence has not yet receive academic attention despite the importance of smartphones in their daily academic and personal lives. This study examined the extent to which students' intelligence dimensions can be predicted using five streams of behavioral information collected from their smartphones. Participants involved 426 students recruited from four high institutions of learning in Sokoto State, Nigeria. Employing a machine learning approach, we successfully predicted students' multiple intelligence based on their smartphone activities collected over a period of 30 days (357,840 log data). Results from our predictive algorithm indicated the possibility of identifying students' intelligences using digital footprints collected from their daily smartphone activities, classified as: interaction behavior; device use behavior; app use behavior; connectivity behavior; and file storage event. While proposing the possibility of these smartphone activities to change in proportion to change in intelligence behavior, we recommend that smartphone manufacturers should improve the digital wellbeing tools in all smartphones to enable tracking of more streams of activities that will ease the prediction of emerging intelligence dimension without having to rely on third party applications.
To connect or not to connect: Visitor preferences for Wi-Fi and cellular network service at a national park
Article
Full-text available
  • Jul 2021
Worldwide mobile devices have become ubiquitous including for visitors to parks and protected areas. Wireless communication technologies, such as Wi-Fi access and cellular network service, are also common but not all parks and protected areas provide these services. Wireless communication technologies could alter the visitor experience, however little is known about visitor preferences for Wi-Fi access and cellular network service nor about experiential outcomes of mobile device usage in natural settings. Therefore, the purpose of this research was twofold: to understand visitor preferences for Wi-Fi access and cellular network service, and experiential outcomes of mobile device usage. The study site for this research was Shenandoah National Park (n = 1442). Using k-means clustering analysis, data were clustered into visitor groups and statistically compared to gain insight into patterns of responses. Visitors mostly reported neutral preferences for Wi-Fi access and cellular network service except for cell service park-wide in all national parks was reported as important. One group of visitors (n = 239) reported that these technologies were unimportant and mobile device usage diminishes their experience. Analysis was also conducted to investigate preferences across age, which showed older demographics reported both cell service park-wide and cell service park-wide in all national parks as important. This information is relevant for management decisions regarding wireless communication technologies access and avail- ability within parks and protected areas.
Mobile Internet usage and usage‐based pricing
Article
Using data on mobile Internet usage of thousands of individuals, we provide some of the first analyses linking mobile usage to key demographics such as income. We find a reverse‐U relationship between mobile Internet usage and income—notably different than the monotonically declining relationship found on home devices. This pattern suggests that data caps are particularly binding on low‐income users. We then construct a simple model of mobile Internet usage that incorporates demand features suggested by our empirical finding and prior empirical findings on device adoption and usage. After abstracting away from cost and two‐sided market considerations, we solve the model, and through comparative statics, identify demand conditions for which usage‐based pricing (via a data cap) is, or is not, revenue enhancing. Key insights from this analysis are: (1) the tendency toward demand‐driven price discrimination is hill‐shaped (increasing then decreasing) in the number of low‐income users and eventually declining in high‐income users, and (2) a relatively high (low) proportion of total value attained from relatively low usage levels by low‐income (high‐income) users increases the tendency toward demand‐driven price discrimination. Hence, to the extent that a market has a large number of low‐income users and/or high‐income users attain a higher proportion of their total value at the cap than low‐income users, the use of caps by providers is more likely driven by cost (or other non‐demand‐side) considerations than by revenue enhancement. Lastly, additional analysis shows a largely monotonically increasing relationship between income and usage intensity (measured as page views in a session), suggesting that, ceteris paribus, price discrimination strategies may be more effective if tied to usage intensity rather than duration.
Context-based Automated Responses of Unavailability in Mobile Messaging
Article
Full-text available
  • Jun 2021
People are not always able to respond immediately to incoming messages on their mobile devices, either due to engagement in another task or simply because the moment is inconvenient for them. This delay in responding could affect social relationships, as there are often expectations associated with mobile messaging and people may experience a lingering pressure to attend to their messages. In this work, we investigate an approach for generating automated contextual responses on behalf of message recipients when they are not available to respond. We first identify several types of contextual information that can be obtained from a user’s smartphone and explore whether those can be used to explain unavailability. We then assess users’ perception of the usefulness of these sensor-based categories and their level of comfort with sharing such information through a Mechanical Turk survey study. Our results show emergent groups with varying preferences with regards to the usefulness and comfort in sharing two types of contextual information: user state and device state. Further, we also observed a strong influence of message context (i.e., message urgency and social tie strength) in users’ perceptions of these auto-generated messages. Our research provides understanding of users’ perceptions of sharing context through an autonomous agent that can help design and create effective approaches towards enabling communication awareness.
Predicting Depression From Smartphone Behavioral Markers Using Machine Learning Methods, Hyperparameter Optimization, and Feature Importance Analysis: Exploratory Study
Article
Full-text available
  • Jul 2021
Background Depression is a prevalent mental health challenge. Current depression assessment methods using self-reported and clinician-administered questionnaires have limitations. Instrumenting smartphones to passively and continuously collect moment-by-moment data sets to quantify human behaviors has the potential to augment current depression assessment methods for early diagnosis, scalable, and longitudinal monitoring of depression. Objective The objective of this study was to investigate the feasibility of predicting depression with human behaviors quantified from smartphone data sets, and to identify behaviors that can influence depression. Methods Smartphone data sets and self-reported 8-item Patient Health Questionnaire (PHQ-8) depression assessments were collected from 629 participants in an exploratory longitudinal study over an average of 22.1 days (SD 17.90; range 8-86). We quantified 22 regularity, entropy, and SD behavioral markers from the smartphone data. We explored the relationship between the behavioral features and depression using correlation and bivariate linear mixed models (LMMs). We leveraged 5 supervised machine learning (ML) algorithms with hyperparameter optimization, nested cross-validation, and imbalanced data handling to predict depression. Finally, with the permutation importance method, we identified influential behavioral markers in predicting depression. Results Of the 629 participants from at least 56 countries, 69 (10.97%) were females, 546 (86.8%) were males, and 14 (2.2%) were nonbinary. Participants’ age distribution is as follows: 73/629 (11.6%) were aged between 18 and 24, 204/629 (32.4%) were aged between 25 and 34, 156/629 (24.8%) were aged between 35 and 44, 166/629 (26.4%) were aged between 45 and 64, and 30/629 (4.8%) were aged 65 years and over. Of the 1374 PHQ-8 assessments, 1143 (83.19%) responses were nondepressed scores (PHQ-8 score <10), while 231 (16.81%) were depressed scores (PHQ-8 score ≥10), as identified based on PHQ-8 cut-off. A significant positive Pearson correlation was found between screen status–normalized entropy and depression (r=0.14, P<.001). LMM demonstrates an intraclass correlation of 0.7584 and a significant positive association between screen status–normalized entropy and depression (β=.48, P=.03). The best ML algorithms achieved the following metrics: precision, 85.55%-92.51%; recall, 92.19%-95.56%; F1, 88.73%-94.00%; area under the curve receiver operating characteristic, 94.69%-99.06%; Cohen κ, 86.61%-92.90%; and accuracy, 96.44%-98.14%. Including age group and gender as predictors improved the ML performances. Screen and internet connectivity features were the most influential in predicting depression. Conclusions Our findings demonstrate that behavioral markers indicative of depression can be unobtrusively identified from smartphone sensors’ data. Traditional assessment of depression can be augmented with behavioral markers from smartphones for depression diagnosis and monitoring.
Menthal: A Framework for Mobile Data Collection and Analysis
Conference Paper
Full-text available
  • Sep 2016
We present a framework for collecting and analyzing mobile users' data. We have built our own framework in order to fulfill specific tracking requirements such as app usage and affect/mood recording. We incentivize the users to share their data through giving feedback. The UX and the interface are primary concerns in our Android app. To be able to process large amounts of data we developed our analysis platform using a modern big data stack. The framework has been deployed in a large scale longitudinal experiment since January 2014. It has attracted more than 400, 000 app installs and more than 350, 000 registered participants.
Menthal - Running a Science Project as a Start-Up
Conference Paper
Full-text available
  • May 2016
We present a software framework for collecting and analyzing mobile users' data. It is used in one of the largest in-the-wild smartphone studies. We attracted a large number of participants by running the study in a start-up format: building a user desirable product and then marketing it through media outlets. From January 2014 until December 2015 we have drawn more than 380 000 app installs with more than 300 000 registered participants and more than 95 000 questionnaire answers. From these, we have collected general phone measurements but also other appealing data such as mood/affect measurements, and big five personality dimensions. In this paper we shortly present our framework and strategy. Then we present general demographic data of a sample of 5 000 participants. In the conclusion section we propose more methods of analysis, how to attract attention to smartphone overuse and how to reduce it.
Understanding Human-Smartphone Concerns: A Study of Battery Life
Conference Paper
Full-text available
  • Jun 2011
This paper presents a large, 4-week study of more than 4000 people to assess their smartphone charging habits to identify timeslots suitable for opportunistic data uploading and power intensive operations on such devices, as well as opportunities to provide interventions to support better charging behavior. The paper provides an overview of our study and how it was conducted using an online appstore as a software deployment mechanism, and what battery information was collected. We then describe how people charge their smartphones, the implications on battery life and energy usage, and discuss how to improve users’ experience with battery life.
Falling asleep with Angry Birds, Facebook and Kindle: A large scale study on mobile application usage
Conference Paper
Full-text available
  • Aug 2011
While applications for mobile devices have become extremely important in the last few years, little public information exists on mobile application usage behavior. We describe a large-scale deployment-based research study that logged detailed application usage information from over 4,100 users of Android-powered mobile devices. We present two types of results from analyzing this data: basic descriptive statistics and contextual descriptive statistics. In the case of the former, we find that the average session with an application lasts less than a minute, even though users spend almost an hour a day using their phones. Our contextual findings include those related to time of day and location. For instance, we show that news applications are most popular in the morning and games are at night, but communication applications dominate through most of the day. We also find that despite the variety of apps available, communication applications are almost always the first used upon a device's waking from sleep. In addition, we discuss the notion of a virtual application sensor, which we used to collect the data.
Predicting The Next App That You Are Going To Use
Conference Paper
  • Feb 2015
Given the large number of installed apps and the limited screen size of mobile devices, it is often tedious for users to search for the app they want to use. Although some mo-bile OSs provide categorization schemes that enhance the visibility of useful apps among those installed, the emerging category of homescreen apps aims to take one step further by automatically organizing the installed apps in a more intel-ligent and personalized way. In this paper, we study how to improve homescreen apps' usage experience through a pre-diction mechanism that allows to show to users which app she is going to use in the immediate future. The prediction technique is based on a set of features representing the real-time spatiotemporal contexts sensed by the homescreen app. We model the prediction of the next app as a classification problem and propose an effective personalized method to solve it that takes full advantage of human-engineered fea-tures and automatically derived features. Furthermore, we study how to solve the two naturally associated cold-start problems: app cold-start and user cold-start. We conduct large-scale experiments on log data obtained from Yahoo Aviate, showing that our approach can accurately predict the next app that a person is going to use.
Device Analyzer
Article
New hardware capabilities and integrated application market stores have created complex usage patterns in modern smart phones. Understanding these patterns is important to both smart phone providers and their customers. Hardware manufacturers might make use of this data when considering design trade-offs of functionality against cost and form-factor. Summarised data of handset use is also useful for device owners trying to stay within data limits on mobile networks or even to select the best tariff. The idea of personal analytics takes this to an extreme in which one collects detailed statistics of many life aspects. Our Device Analyzer application aims to collect a largescale research data-set of phone usage. Device Analyzer gathers data about running processes, wireless connectivity, the phone’s location, GSM communication, battery state and a
AppAware: which mobile applications are hot?
Conference Paper
  • Jan 2010
Today most mobile operating systems provide users with an application portal where they can search for applications published by third-party developers. However, finding new apps is not an easy task and requires either to know what to look for or to go through an endless list of applications. In this paper we present work in progress of a platform that allows its users to discover mobile applications in a serendipitous manner. AppAware is a mobile application that captures and shares installations, updates, and removals of Android programs in real time. Accordingly, AppAware allows its users to see what applications are being installed right now or around their position by other people, thus introducing a new way of interaction with application portals and other mobile users.
Device Analyzer: Large-scale mobile data collection
  • Jan 2014
  • Perform Eval Rev
  • 53-56
  • T Daniel
  • Andrew Wagner
  • Alastair R Rice
  • Beresford
Daniel T Wagner, Andrew Rice, and Alastair R Beresford. 2014. Device Analyzer: Large-scale mobile data collection. ACM SIGMETRICS Performance Evaluation Review 41, 4 (2014), 53-56.
  • Jan 2016
  • Ericsson
Ericsson. 2016. Ericsson Mobility Report. (2016).
Menthal -Running a Science Project as a Start-Up
  • Jan 2016
  • Konrad Ionut Andone
  • Mark Błaszkiewicz
  • Boris Eibes
  • Christian Trendafilov
  • Alexander Montag
  • Markowetz
Ionut Andone, Konrad Błaszkiewicz, Mark Eibes, Boris Trendafilov, Christian Montag, and Alexander Markowetz. 2016a. Menthal -Running a Science Project as a Start-Up. In Computing in Mental Health, Workshop at CHI 2016. ACM.