Dataset Information

Temporal and cultural limits of privacy in smartphone app usage.

ABSTRACT: Large-scale collection of human behavioural data by companies raises serious privacy concerns. We show that behaviour captured in the form of application usage data collected from smartphones is highly unique even in large datasets encompassing millions of individuals. This makes behaviour-based re-identification of users across datasets possible. We study 12 months of data from 3.5 million people from 33 countries and show that although four apps are enough to uniquely re-identify 91.2% of individuals using a simple strategy based on public information, there are considerable seasonal and cultural variations in re-identification rates. We find that people have more unique app-fingerprints during summer months making it easier to re-identify them. Further, we find significant variations in uniqueness across countries, and reveal that American users are the easiest to re-identify, while Finns have the least unique app-fingerprints. We show that differences across countries can largely be explained by two characteristics of the country specific app-ecosystems: the popularity distribution and the size of app-fingerprints. Our work highlights problems with current policies intended to protect user privacy and emphasizes that policies cannot directly be ported between countries. We anticipate this will nuance the discussion around re-identifiability in digital datasets and improve digital privacy.

SUBMITTER: Sekara V

PROVIDER: S-EPMC7887199 | biostudies-literature | 2021 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Temporal and cultural limits of privacy in smartphone app usage.

Sekara Vedran V Alessandretti Laura L Mones Enys E Jonsson Håkan H

Scientific reports 20210216 1

Large-scale collection of human behavioural data by companies raises serious privacy concerns. We show that behaviour captured in the form of application usage data collected from smartphones is highly unique even in large datasets encompassing millions of individuals. This makes behaviour-based re-identification of users across datasets possible. We study 12 months of data from 3.5 million people from 33 countries and show that although four apps are enough to uniquely re-identify 91.2% of indi ...[more]

PMID: 33594096

Similar Datasets

Project description:BACKGROUND:Advancements in mobile phone technology offer huge potential for enhancing the timely delivery of health behavior change interventions. The development of smartphone-based health interventions (apps) is a rapidly growing field of research, yet there have been few longitudinal examinations of how people experience and use these apps within their day-to-day routines, particularly within the context of a hybrid Web- and app-based intervention. OBJECTIVE:This study used an in-depth mixed-methods design to examine individual variation in (1) impact on self-reported goal engagement (ie, motivation, self-efficacy, awareness, effort, achievement) of access to a weight management app (POWeR Tracker) when provided alongside a Web-based weight management intervention (POWeR) and (2) usage and views of POWeR Tracker. METHODS:Thirteen adults were provided access to POWeR and were monitored over a 4-week period. Access to POWeR Tracker was provided in 2 alternate weeks (ie, weeks 1 and 3 or weeks 2 and 4). Participants' goal engagement was measured daily via self-report. Mixed effects models were used to examine change in goal engagement between the weeks when POWeR Tracker was and was not available and whether the extent of change in goal engagement varied between individual participants. Usage of POWeR and POWeR Tracker was automatically recorded for each participant. Telephone interviews were conducted and analyzed using inductive thematic analysis to further explore participants' experiences using POWeR and POWeR Tracker. RESULTS:Access to POWeR Tracker was associated with a significant increase in participants' awareness of their eating (?1=0.31, P=.04) and physical activity goals (?1=0.28, P=.03). The level of increase varied between individual participants. Usage data showed that participants used the POWeR website for similar amounts of time during the weeks when POWeR Tracker was (mean 29 minutes, SD 31 minutes) and was not available (mean 27 minutes, SD 33 minutes). POWeR Tracker was mostly accessed in short bursts (mean 3 minutes, SD 2 minutes) during convenient moments or moments when participants deemed the intervention content most relevant. The qualitative data indicated that nearly all participants agreed that it was more convenient to access information on-the-go via their mobiles compared to a computer. However, participants varied in their views and usage of the Web- versus app-based components and the informational versus tracking tools provided by POWeR Tracker. CONCLUSIONS:This study provides evidence that smartphones have the potential to improve individuals' engagement with their health-related goals when used as a supplement to an existing online intervention. The perceived convenience of mobile access to information does not appear to deter use of Web-based interventions or strengthen the impact of app access on goal engagement. A mixed-methods design enabled exploration of individual variation in daily usage of the app-based tools.

Project description:BackgroundSocial communication via instant messaging (IM) and social networking (SN) apps makes up a large part of the time that smartphone users spend on their devices. Previous research has indicated that the excessive use of these apps is positively associated with problematic smartphone use behaviors. In particular, image-based SN apps, such as Instagram (Facebook Inc) and Snapchat (Snap Inc), have been shown to exert stronger detrimental effects than those exerted by traditional apps, such as Facebook (Facebook Inc) and Twitter (Twitter Inc).ObjectiveIn this study, we investigated the correlation between individuals' tendencies toward smartphone use disorder (SmUD) and objective measures of the frequency of smartphone usage. Additionally, we put to test the hypothesis that the pathway linking the frequency of actual smartphone usage to self-reported tendencies toward SmUD was mediated by the increased frequency of IM and SN app usage.MethodsWe recruited a sample of 124 adult smartphone users (females: 78/124, 62.9%; age: mean 23.84 years, SD 8.29 years) and collected objective information about the frequency of smartphone and SN app usage over 1 week. Participants also filled in a self-report measure for assessing the multiple components of tendencies toward SmUD. Bivariate associations were investigated by using Spearman correlation analyses. A parallel mediation analysis was conducted via multiple regression analysis.ResultsThe frequency of smartphone usage, as well as the use of IM apps (Messenger, Telegram, and WhatsApp [Facebook Inc]), Facebook, and image-based apps (Instagram and Snapchat), had significant positive associations with at least 1 component of SmUD, and the cyberspace-oriented relationships factor exhibited the strongest associations overall. We found support for an indirect effect that linked actual smartphone usage to SmUD tendencies via the frequency of the use of image-based SN apps.ConclusionsOur novel results shed light on the factors that promote SmUD tendencies and essentially indicate that image-based SN apps seem to be more strongly associated with problematic smartphone behaviors compared to IM apps and traditional SN apps, such as Facebook.

Project description:IntroductionAlthough there is evidence linking the relationships between smartphone usage with health, stress, and academic performance, there is still inadequate knowledge about the influence on pro-environmental behaviors. This study seeks to bridge this gap by adapting the theory of attribution framework to examine the effects of personal norms, social norms, perceived behavioral control on pro-environmental behavior of smartphone usage in children.MethodsA total of 225 children aged between 11 to 12 from eight selected public primary schools at the Hsinchu Science and Industrial Park in Taiwan were surveyed. Two distinct groups (excessive versus moderate usage) were purposefully selected for comparison, of which 96 participants were excessive smartphone users while the remaining 129 were moderate smartphone users.ResultsFindings revealed significant differences between excessive and moderate smartphone usage children groups in personal norms (p < 0.001), social norms (p = 0.002), perceived behavioral control (p = 0.001), and pro-environmental behavior (p = 0.001). Findings for excessive smartphone usage children showed that social norms (β = 0.428, t = 4.096***, p < 0.001) had a direct predictive impact on pro-environmental behavior. In contrast, while there was no direct path established between personal norms and pro-environmental behavior (β = 0.177, t = 1.580, p > 0.05), as well as social norms and pro-environmental behavior for moderate smartphone usage children (β = 0.181, t = 1.924, p > 0.05), but such a relationship could be developed through the mediating effect of perceived behavioral control (β = 0.497, t = 4.471***, p < 0.001).DiscussionThe results suggested that excessive smartphone usage children lack positive perceived behavioral control, and their pro-environmental behavior could only be predicted through explicit social norms, whereas pro-environmental behavior of moderate smartphone usage children was implicitly influenced by personal norms through perceived behavioral control.

Project description:BackgroundPreventing progression from prediabetes to diabetes-or slowing the progression of diabetes-is an urgent task worldwide. Previous studies have shown that mobile health (mHealth) may powerfully support self-management for patients with type 2 diabetes. Certainly, mHealth improves health care efficiency and gives patients convenient access to self-management of their own health. Many health care apps are available right now, and their use in clinical studies with large-scale real-life data is expected. However, the usage patterns of those apps-especially in the absence of intervention by medical professionals-remain unknown.ObjectiveWe developed GlucoNote, an app that uses Apple's ResearchKit to support self-management for patients with type 2 diabetes and prediabetes; the app does not require prescription or intervention by medical professionals. We evaluated its usage patterns via a remotely conducted study.MethodsiPhone users across Japan who have type 2 diabetes or prediabetes were free to download GlucoNote and to participate in the study after they provided consent electronically on the app. The 522 users who enrolled in the study within 1 year of its release were analyzed. We analyzed the retention rates of 357 participants who recorded at least 1 of 4 items-body weight, blood sugar, blood pressure, or dietary information. Characteristics of participants who used GlucoNote longer than 4 weeks (robust users) were compared with those of participants who did not (nonrobust users). The changes among robust users were evaluated.ResultsThe median observation and retention durations were 382 days (interquartile range [IQR] 275-423) and 8 days (IQR 1-63), respectively. The retention rates for 2 days and for 4, 8, and 12 weeks were 0.627 (95% CI 0.575-0.675), 0.353 (0.304-0.403), 0.272 (0.227-0.319), and 0.220 (0.179-0.265), respectively. Men were more likely to be robust users than women (P=.02). At week 0, robust users were more likely than nonrobust users to have a higher daily energy intake (median 1595 [IQR 1198-1788] kcal vs 1451 [IQR 769-1657] kcal; P=.04) and have higher daily step counts (median 6108 [IQR 3797-9227] vs 5171 [IQR 2885-7258]; P=.001). Among robust users, body weight decreased from weeks 0 to 4 (mean 71.3 [SD 14.1] kg to 70.8 [SD 13.9] kg; P=.002) by mean 0.6% (SD 1.6).ConclusionsGlucoNote offered a valuable opportunity to evaluate usage patterns of apps. Future challenges include improving low retention rates and evaluating their effects.

Dataset Information

Temporal and cultural limits of privacy in smartphone app usage.

Publications

Temporal and cultural limits of privacy in smartphone app usage.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets