6. Resultados y discusión
6.2.2 Creación del modelo y estudio de la exactitud
2.6.3.1 Visualisation for Online Users’ Profile
Many studies have employed visualisation techniques as a way of presenting users’ profiles due to their simplicity in conveying information [114]. Middleton et al. [115] developed a system for searching paper databases and online paper recommendation . The system offers profile visualisation for users that displays their interests and enables the collection of further inform- ation on their interests. User interest profiles are derived and updated daily and visualised to users as a time/interest graph. Heer and Boyd [116] presented a visualisation system that en- ables end-users to explore and analyse large-scale social networks by simplifying the discovery of their online communities and awareness of information exposure. It enables the accessing and searching of users’ profiles on an online dating site via visualisation, and delivers the so- cial network based on a common node-link layout, where nodes are the users and links are the friendships. Evaluation revealed that users were able to explore their online community while enjoying the experience, and they used the system’s features effectively. Similarly, Tchuente et al. [117] introduced a method for presenting temporal graph visualisation of users’ short and long-term interests that considers the evolving interests of users. A third-party application was implemented using the Facebook API which 85 users installed in their profiles. The derived in- terests were projected as a 3-D co-occurrences matrix where the level of temporal granularity of their interests can be specified in various periods. Furthermore, Plumbaum et al.[118] developed a user-centric system with a personal user interface for aggregating users’ profiles from various web applications, and enabling users to manage and share their personal information in a pri- vacy preserving environment. The personal user interface visualises users’ personal profiles and allows them to track their personal information in different applications as well as control their data sharing. Church et al. [119] conducted a study investigating the impact of the visualisation type used in a LBS search interface on users’ experience of information discovery . They found having both map- and text-based user interfaces is more effective in providing the user with the required information and enhancing their experience. More recently, Cuttone et al.[120]
2.6 Privacy-Enhancing Technologies for Privacy Awareness 35
implemented a personal informatics tool for Android smart phones that offers interactive visual projections of personal information on users’ mobility and social interactions. They introduced a visual spiral timeline that displays the mobility patterns across various temporal periods. The social interactions are captured by monitoring the detected devices, where they are grouped in the form of bins depending on their timestamp, and each user is assigned a weight based on their meeting frequency. Initial results from users’ activity log files denote that they are able to know more about their behavioural patterns . In addition, Vosecky et al. [10] modelled users’ geographical interests shared on Twitter in relation with their corresponding disclosed locations by developing an interactive visualisation system. The system presents the users’ geographical topics integrated with recommendations in many visualisations, including user-based by show- ing an individual’s user interest profile and news recommendations, region based by showing topics and users associated with a certain region, and topic based by presenting the terms related with a particular topic.
Generally, visualising location data facilitates perception of the information presented. We used visualisation methods in this work as a means of representing a users’ location-based profile, hence enhancing their awareness of their shared data.
2.6.3.2 Visualisation for Privacy Awareness
Visualisation of privacy warnings was found to be effective in increasing user awareness of privacy implications [101]. Studies employed these techniques as a way of providing feedback on users’ information accessibility based on the user specified privacy preferences. Anwar and Fong [33] developed a visualisation tool that enables users to explore how their profiles are viewed from the user’s perspective in their social connections in order to provide means for understating the privacy implications of the access controls. Participants were asked to perform policy analysis with and without the tool. As a result, the participants were able to perform policy assessment more accurately when using the visualisation tool . In addition, Rode et al. [121] explored visualisation techniques to present system activities and integrate configuration with action as a way of increasing users’ understating of the consequences of their action with a system. They implemented a prototype that represents the client’s shared workplace as an interactive pie-shaped interface where each slice corresponds to a user’s space. An initial user study revealed that the participants were easily able to set permissions and found that the privacy level is intuitive. They were also able to understand the activities broadcasted on the interface. Similarly, Wang et al.[122] implemented an interactive visualisation system that aids users in specifying sharing preferences for their personality traits that are extracted from Twitter. This tool allows users to understand their derived personality traits by showing three types of traits as labelled bars, where each is linked to further sub-traits and coloured distinctively. The filled
36 2.6 Privacy-Enhancing Technologies for Privacy Awareness
bar’s length indicates the score for the extracted trait. Evaluation revealed that participants found the visualisation tool useful in configuring their settings.
Several studies have focused on raising users’ awareness of collection of their online data by web services. The concept of online interactive privacy was introduced by Kani-Zabihi and Helmhout [123]. This is concerned with any tool or user interface that supports privacy aware- ness and understanding of online privacy threats by showing the personal information flow. A user study was conducted using a prototype in the form of a mock-up council where participants were asked to perform pre-defined tasks. As a result, online interactive features increase users’ privacy awareness and motivate them to learn more about how their personal information is used. Angulo et al. [124] developed a tool that visualises online data disclosure with the aim of supporting usable data transparency. It is basically a user interface that shows an overview of users’ disclosed data to various online services and enables them to access their data col- lected on the service side. The prototype represents a user as a profile picture in the centre of the interface that is connected with the attributes of their disclosed data at the top, and with service providers that data are resealed to at the bottom. Scenario-based usability testing and a workshop session were conducted, which revealed improvement to users’ awareness of their data disclosure to web services.
Focussing on GeoSocial environments, Brush, Krumm and Scott [57] studied users’ attitudes towards their location privacy when sharing and tracking long-term location information. They collected location traces of 32 participants using GPSs over a period of two months, and then showed them personal visualised maps of their location tracks using five different obfuscation methods. They argued that some of the participants choices of obfuscation methods were not compatible with their privacy concerns, and that this might be due to participants’ lack of un- derstanding of the implications of obfuscation. More recently, Tang, Hong and Siewiorek [92] developed three types of isomorphic visualisations: text-, map-, and time-based, that also con- sider spatial and temporal properties of sharing historical location, such as the physical location and duration, and included four place labels categories (geographical or semantic, general or specific). They applied these three visualisations to their participants’ collected GPS data and then showed it to them. The majority of the participants mentioned concerns related to their physical privacy when showing them their visualised location history. Consequently, they all preferred text-based visualisation, since they believed it to reveal the least location information and to be the least appealing to others because it requires more effort to understand.
Visualisation methods have shown to effectively serve privacy-oriented purposes. Yet, utilising and studying the impact of visualisation on location privacy awareness, particularly in GeoSNs, is yet to be investigated. Thus, this work employs visualisation techniques within the proposed location privacy awareness solutions that aim to present the privacy implications of location