Usability Study of a Semantic User Model Visualization for Social Networks Fedor Bakalov and Birgitta König-Ries Tobias Hennig and Gabriele Schade Friedrich Schiller University of Jena Erfurt University of Applied Sciences fedor.bakalov@uni-jena.de tobias_hennig@hotmail.de birgitta.koenig-ries@uni-jena.de schade@fh-erfurt.de ABSTRACT of friends she is most interested in will appear more prominently In this paper, we report on a usability study of the Introspective- than the ones from friends the user does not have close contact to. Views interface for visualizing semantic user models in social net- works. In the first part of this paper, we describe how the inter- User models play an essential role for achieving such personaliza- face can be used in a social network for visualizing and editing tion effects. In a social network, they can provide such informa- ontology-based user interest models. In the second part, we de- tion as what friends and groups the user is most interested in, what scribe the study we conducted to evaluate the usability of this vi- games and applications she likes, and what her general interests sualization. We describe the results of four evaluation methods: are (e.g., meeting new people, keeping in touch with classmates, task-based experiment, eye tracking, interviewing, and a question- job search, and so on). Some of this information can be explic- naire. itly provided by the user, whereas some is inferred by the system automatically based on the user’s interaction. The contributions of this paper are three-fold: First, it describes an application of the IntrospectiveViews interface to user model visu- In many cases, the system-generated user models are hidden from alization in social networks. Second, it describes and exemplifies the user. This, however, may lead to grave usability problems and a methodology to usability evaluation of interactive visualizations may well cause the user to not accept the system. For instance, of user models. Third, it reports on the outcomes of such an eval- it violates two of Nielsen’s ten usability principles [16]: Hiding uation: It identifies what visual techniques and interaction patterns user models occludes the system status and hinders control on the users deem to be usable and attractive for working with models of personalization, which may lead to errors, e.g. issuing irrelevant their interests, but it also reports on the crucial usability problems recommendations. of the interface and describes possible solutions to these problems. We believe that the latter two contributions to be of interest for In order to avoid the above mentioned problems, user models need researchers working on similar interfaces for visualization of user to be scrutable. This means that the user needs to be able to view models. and adapt the information contained in her user model [10]. Jame- son [9] argued that allowing inspection and parametrization of user models are important measures to achieve predictability, transpa- Keywords rency, and controllability of an adaptive system. According to Cook usability, information visualization, scrutable user models, ontology- and Kay [5], the user needs to be able to understand the provenance based user models, social networks of information in her user model, e.g., the user needs to understand why the system believes she is interested in a certain topic or a 1. INTRODUCTION certain entity. Finally, Orwant [17] argued that scrutability is an Nowadays, a number of social networks personalize their content essential step towards establishing trust between the user and an and navigation based on the information about individual users and adaptive system. groups. Personalization effects in a social network can take various forms. This can be, for example, the recommendation to connect However, opening the system-generated user models may be very to other users you may know or jobs you may be interested in. In challenging in case of complex and large models. For instance, in a addition to that, the personalization can take the form of content semantic ontology-based user interest model the items can be inter- filtering or sorting. For instance, to help the user to keep updated connected through a number of semantic relations, which in their with the activities of her close friends, the friends’ posts can be turn can be used for propagating interest among the items to com- sorted or filtered based on the user’s individual interests – the posts pensate the scarcity of information about the user. For an average user it may be difficult to comprehend these semantic relations and their effects on the interest propagation and the end personaliza- tion effect. The large models containing a big amount of items and relations make it even more difficult. In this paper, we describe the application of the IntrospectiveViews interface presented in our prior work [2, 3] for visualizing and edit- Workshop on Visual Interfaces to the Social and Semantic Web ing semantic user models in a social network. Also, we report the (VISSW2011), Co-located with ACM IUI 2011, Feb 13, 2011, Palo Alto, results of a thorough usability study of this interface that we con- US. Copyright is held by the author/owner(s). ducted considering its use in a social network. The contributions relations among them on spirals in a 2D space. However, unlike the of this paper are three-fold: First, it describes an application of IntrospectiveViews described in this paper, both interfaces support the IntrospectiveViews interface to user model visualization in so- only visualization of user models. They do not address the issue of cial networks. Second, it describes and exemplifies a methodology providing the user edit access to her model. Another distinguishing to usability evaluation of interactive visualizations of user models. feature of IntrospectiveViews is its ability to visualize ontology- Third, it reports on the outcomes of such an evaluation: It identifies based user models providing rich semantics about user interests. what visual techniques and interaction patterns users deem to be usable and attractive for working with models of their interests, but 3. DESCRIPTION OF THE INTERFACE it also reports on the crucial usability problems of the interface and In this section, we describe the application of IntrospectiveViews describes possible solutions to these problems. We believe that the for visualizing an ontology-based user interest model in a social latter two contributions to be of interest for researchers working on network. This model provides information about items users in a similar interfaces for visualization of user models. social network can be interested in, including people, groups, or- ganizations, countries, cities, and topics. The model is represented The rest of this paper is organized as follows: Sect. 2 provides as an overlay model, i.e., user interests are defined as a weighted a short overview of the previous research related to visualization overlay of the ontology instances represented in the domain model. of user models. In Sect. 3, we describe how IntrospectiveViews The user model is defined as a set of tuples (U, T, I, V, L), where: can be used for visualizing and editing semantic user models in a social network. In Sect. 4, we elaborate on the usability study of the interface, which included four methods: task-based exper- • U - user ID iment, eye tracking, interviewing, and a questionnaire. Then in Sect. 5 we summarize the most critical usability problems identi- • T - instance of a term from the domain ontology fied in the evaluation and outline a number of possible solutions to these problems. Finally, Sect. 6 provides a summary of this paper • I - degree of interest and an outlook into the future work. • V - validity of interests • L - ID of the last update log entry (reference to the user model 2. RELATED WORK log) A number of approaches have been proposed to visualizing scrutable user models. PeerGlass architecture [13] provides a visual method to exploring user models through a Rolodex of model planes, where The semantic description of instances is defined in the domain on- each plane represents a certain type of user interests. The um_view tology formalized in OWL3 . For each instance, the ontology pro- interface [5] allows traversing through a user model by expanding vides such information as the instance’s class, attributes, and se- the tree of leaves and viewing detailed information about each item mantic relations to other instances. For more information on the in the model. VlUM [1] and its successor SIV [11] are capable of user model and domain ontology refer to our prior work [4]. visualizing large user models and enable users to get an overview of the whole model, view a subset of related beliefs, filter items In the IntrospectiveViews interface, this user interest model is vi- by relevance, and obtain detailed information about the displayed sualized as a collection of keywords displayed on a circular sur- items. STyLE-OLM [6] and Flexi-OLM [14] visualize open learner face consisting of three colored rings each representing an interest models using concept graphs and trees respectively. group (Fig. 14 ). The color scheme of rings is chosen according to the hot-and-cold metaphor, where hot, represented by red, denotes Also, the literature reports on a number of approaches to visual- interest and cold, represented by blue, denotes no interest. The col- izing the structures of online communities for end-users. For in- ors between red and blue denote partial interest. The positioning stance, Vizister [7] visualizes the user’s community as a graph con- of keywords is determined by the exact degree of interest, i.e., the sisting of egocentric networks and allows the user to explore it in a closer a keyword appears to the center, the higher interest it repre- playful manner. A novel metaphor and rich-interaction patterns for sents. The interest is also encoded into the font size of items. In exploring social networks are implemented in Fidg’t Visualizer1 : addition to that, the items are grouped into circular sectors by type, the interface visualizes the relations among members of the net- i.e., the ontology class they belong to (e.g. person, group, country, work and their tags using the metaphor of magnetism. PieSpy [15] etc.). visually displays the network’s structure based on the relations in- ferred from an Internet Relay Chat (IRC). TouchGraph2 provides a IntrospectiveViews follows Shneiderman’s visual information seek- number of visual representations of LiveJournal’s communities. ing mantra [18]: “overview first, zoom and filter, then details-on- demand”. It offers users an overview of all items present in their However, there has been little work done in the area of user model interest model, it allows zooming into different parts of the model, visualization for social networks. Tchuente et al. [19] describe an filtering keywords according to different criteria, and it will provide approach to visualizing short- and long-term user interests in online details on a specific item upon request. Let us take a closer look on social networks. The proposed visualization displays user interests, how this is achieved. i.e., friends and topics, as a graph of interconnected items. The items are displayed in rectangles denoting the temporal aspect of The entire collection of interests can be zoomed in and out using ei- user interests. Kim [12] describes an application of the VUDM ther the zoom slider present in the Navigator window or the mouse (Visual User model Data Mining) tool for visualizing user models wheel. This feature allows quickly switching from the overview to in social networks. The tool visualizes users, groups, topics, and 3 Web Ontology Language - http://www.w3.org/TR/owl-features/ 1 http://www.fidgt.com/visualize 4 Screenshots and screencasts available at http://www.minerva- 2 http://touchgraph.com portals.de/research/introspective-views/ Figure 1: IntrospectiveViews the detailed view. The last can be especially useful in large mod- els consisting of hundreds items. The user can navigate through the collection (e.g. in an enlarged view) by dragging the surface in the corresponding direction. The interface supports two filter- ing options. Items can be filtered by type and by interest degree. For instance, the user can display only groups and people that she is most interested in by selecting the corresponding checkboxes in the Interest Groups and Types windows located on the right side. Also, the user can obtain additional information about items. Right click on an item will display the item’s context menu (Fig. 2), through which the user can access the item’s description, display Figure 2: Item context menu the semantic relations to related items (Fig. 3), and get justifica- tion of her interest. The last explains how the system determined the interest: it can be based on the user’s frequent communication Finally, through the item’s context menu (Fig. 2), the user can set with the entity, propagated from other items using the semantic re- the item’s privacy and personalization settings. This provides the lations, and specified by the user explicitly through Introspective- user a better control on her privacy and the personalization effects. Views. Through the context menu, the user can also quickly ac- For instance, the user can define which interests may be used for cess the information relevant to the selected items. For instance, by which personalization effects. clicking a friend’s name, the user can access the chat history, posts, pictures, and other information posted by the friend on the network. 4. USABILITY STUDY In addition to viewing, the interface enables the user to edit her In our previous study [2] we evaluated the usefulness and usability interest model. The user can change interest degree for a certain of IntrospectiveViews as a generic interface for user model visual- item by simply dragging it on the surface: Moving items closer ization. The results of that study showed which features users deem to the circle center increases the interest and moving towards the important and whether they are sufficiently intuitive and easy-to- edge decreases it. New interests can be easily added by making use. Based on the results, we redesigned the interface. We removed right click in any place on the surface and selecting the Add New the features that were deemed unnecessary or disturbing and re- Interest item from the context menu. Items can be removed by designed the important ones to make them easier to find and use dragging on to the recycle bin or through the item’s context menu. (refer to Table 1 in [2] for the complete list of features). Among eight tasks using the interface. During the completion of tasks, the subject was filmed using audio/video recording equipment. Also, we recorded the participant’s gaze and mouse movement. After- wards, the subject was interviewed and asked to fill out an evalu- ation question. The rest of this section provides a more detailed description of each method and its results. 4.1 Task-Based Experiment In general, task-based experiments are conducted to test how intu- itive an interface is. To measure this, the participants are given spe- cific tasks to solve using the interface. For our experiment we iden- tified eight tasks representing the most common goals that users might want to achieve working with IntrospectiveViews. During Figure 3: Visualization of semantic relations the experiment each user had to complete each of these tasks using the interface. others, we removed the feature for automatic propagation of user Task 1. Acquaint yourself with the interface. Guess what pur- explicit changes, changed the meaning for font size of items (in- pose it has and speak out on it. The goal of this task was to stead of encoding the log frequency, it now encodes the interest determine whether the user can understand the purpose and func- degree), and improved the visual presentation of grouping by type. tion of IntrospectiveViews without any introduction or manual. To describe the first impression, the participants used such words as: The goal of the study described in this paper was to evaluate us- globe, tag-cloud, portal, diagram, eye, and mindmap. Having worked ability of IntrospectiveViews as an interface for visualizing user with the interface for a little, most subjects noticed that items are interest models in social networks. We believe that in the context radially positioned on the circle by interest and grouped into slices of a social network, it is important that the visualization interface by type. Also, they relatively quickly understood the functions of is not only easy-to-use, but also engaging and visually attractive. the three panels located on the right side, i.e., Navigator, Interest Hence, in this study we included not only the techniques that evalu- Groups, and Types. With respect to the purpose of the interface, ate the inherent usability (pragmatic quality), but also the methods the subjects either had no idea at all or expressed themselves very for evaluating the hedonic qualities and the outward appearance. vaguely about it. However, it is important to take into considera- The evaluation consisted of an expert evaluation and a user study. tion that the interface was not launched from a real social network, However, due to the page limit, we describe the user study only. which makes it more difficult to guess its purpose. Nevertheless, For the complete description of the whole study, including the ex- an important observation here is that most participants understood pert evaluation, refer to [8]. the meaning of the way the items are displayed on the surface and noticed the affordability to interact with them. After the subject The user study took place at the Usability Lab of Erfurt University had completed this task, the purpose and possible applications of of Applied Sciences. It was conducted with 10 test persons, 5 male the interface were explained by the evaluator. and 5 female, in age between 20 and 23. All subjects had medium or high level of technical knowledge. While we are aware that this Task 2. Show all persons you are interested in. The goal of the group is rather homogenous, we believe that it represents typical second task was to determine whether the user comprehends the users of many social networks. filtering options of the interface, namely, filter by interest degree and filter by type. After having acquainted themselves with the In the study we used seven techniques, namely task-based exper- interface, most participants completed this task without problems. iment, audio and video recording, mouse and eye tracking, inter- viewing, and a questionnaire. We chose the combination of these Task 3. View privacy settings for an item. Here, we aimed to de- techniques in order to find as many usability problems as possi- termine whether the user can easily find out how to access the pri- ble, to better understand why users experience these problems, and vacy settings of items. Most subjects had problems with this task. to find solutions to them. The task-based experiments reveal how As the first attempt, many clicked the Settings button on the toolbar, successful users can achieve their goals with the interface and what which was supposed (was not implemented at the time of study) to difficulties they experience. Filming, eye tracking, and click stream show the global settings of the system. After failing to complete recording during the task completion help the usability experts to the task that way, almost all participants tried to find the item on perform a more detailed analysis of the user’s interaction. Record- the surface by minimizing the number of items on the screen using ing the user’s eye movement can be especially helpful for answer- the filter by type function. Once the item was found, all participants ing why the user experiences problems while completing a certain except one made right click on the item and selected the appropri- task, e.g., if the user is searching for the required control element on ate menu item from the popped-up context menu (Fig. 2). One a wrong toolbar. Interviews and questionnaires are helpful for get- person tried to access the settings by making double click on the ting the user’s subjective reaction and opinions. They help to find item. From this task, we have learned that performing a task on a what the users like and what they do not like. Also, in interviews specific item might be complicated by the difficulty to find the item the participants can provide valuable insights into how to improve on the surface. the system. Task 4. Find out how the system determined your interest in The course of the study was organized as follows: After a short an item. As the first step, the subjects searched for the given item, introduction, the subject was asked to sign a consent form and fill for which all but two subjects filtered out the items of the irrele- out a personal details form. Then the person was asked to complete vant types and in the reduced number of displayed items found the given one. Then they opened its context menu (Fig. 2) by making that the user could quickly notice and comprehend it. right click. However, in the context menu, only two subjects se- lected the right item, Your Interest. All others could not find it from 4.3 Interviews the first try, but simply were checking all menu items one after an- After the task-based experiment, the subjects were asked to speak other. This means that the Your Interest label is not self-explanatory out on the positive and negative characteristics of Introspective- and understandable, hence should be renamed or else the informa- Views as well as on their suggestions to further improve the in- tion about the user interest should be accessed through some other terface. means. Positive Characteristics. The participants spoke out on the inter- Task 5. Display relations of an item. Most participants suc- face as novel, innovative, cool, and easy-to-use. Many said that it cessfully completed this task by making right click on the given would be a very useful and helpful tool in social networks. With item and selecting the corresponding command from its context respect to the hot-and-cold color metaphor, half of the participants menu. Visualizing semantic relations among items (Fig. 3) was ap- said that it is an appropriate and comprehendible use of colors. The praised by most subjects as a useful feature. However, a number of changing font size of labels on dragging and the highlighting of them complained that the connecting arcs disappear when an item circular sectors on mouse over were also rated as positive. The in- is moved. This means that the interface should provide a better teraction concept for changing the interest degree by dragging the control on visualizing the relations and allow the user to control label on the surface was rated as intuitive and easy-to-use. The vi- not only the turning on, but also the turning off of the lines. sualization of semantic relations among items was rated especially positive, useful, and novel. Task 6. Change interest degree in an item. In Introspective- Views, this task can be completed by dragging items on the surface. Negative Characteristics. One of the most often criticized prob- The task was successfully completed by all participants. Many ap- lems was the overlapping of items, which makes it difficult to read praised the changing font size of items during the dragging. How- and manipulate them. Often, in order to read the names of over- ever, we observed two difficulties with this task. First, for some lapped labels, the participants moved them in different directions, subjects, it took quite a long time to find the necessary item. Sec- which caused unintended change of interest degree. Also, some ond, some subjects had problems to seize the item when it was participants said that the interface has too many control elements overlapped by other items. (especially on the toolbar) and this may confuse the user for the first time. However, another important critique was the poor visi- Task 7. Delete an item. All subjects completed this task success- bility of some control elements: Some participants discovered the fully. Four subjects deleted the item by dragging it onto the recycle context menus of items and the surface by chance after having used bin and the rest did it through the item’s context menu. the interface for a while. In the beginning, it was not clear for them that the items and the surface can be manipulated through the con- Task 8. Add a new item in the model. All subjects completed text menu. this task without serious problems since they had already (during the completion of previous tasks) noticed the Add Interest item in Suggested Improvements. During the interview, the participants the surface’s context menu that can be opened by right click on the made a number of interesting suggestions to further improve the in- surface. Many subjects appraised that the new interests are added terface. One of them was the suggestion to provide a short tutorial in the place from which the context menu was opened. explaining the important functions of the interface. The explana- tion can be provided in the form of callouts displayed next to the 4.2 Eye-Tracking corresponding controls. This tutorial could be automatically dis- During the task-based experiment, the participants’ gaze was recor- played upon the first time with the interface and invoked later upon ded using the eye-tracking equipment. The results of eye-tracking request. Another suggestion was to change the zoom function to show the series of gaze fixations the participants made while look- show different amount of details at different zoom levels, so that in ing at the interface and the density of fixations. Fig. 4 shows a plot a zoomed out view only the most interesting items would be shown, of the series of gaze fixation one participant has made during the but by zooming in the interface would be incrementally showing first ten seconds with the interface. The analysis of gaze fixations more of the less interesting items. Apart from that, it was suggested for all the participants shows, that the user makes the first gazes at to allow the user to define own types and organize the items accord- the labels in the red area of the surface. This is a good result since ing to the self-defined types. Finally, a number of participants said this area contains the items the user is strongly interested in, which that the interface should allow changing the color scheme of the is the most important information here. The next gazes were made circular surface. either at the side windows, e.g., Navigator, or the toolbar. Also, many participants spent a considerable amount of time on reading 4.4 Questionnaire the type labels identifying the circular sectors. To evaluate the attractiveness and joy-of-use of IntrospectiveViews, we used the AttrakDiffTM tool 5 , which includes an online ques- The cumulate gaze density in the first twenty seconds (Fig. 5) shows tionnaire and a number of diagrams visualizing the collected feed- that during this period of time the participants looked more at the back. The AttrakDiffTM questionnaire addresses the subjective at- control elements (toolbar and side windows) than at the labels. tractiveness of a product as a composite characteristic influenced This, however, can be caused by the novelty of the interface and the by four qualities: lack of knowledge about its purpose and function. Having looked at the labels on the surface, the participants could not fully under- • Pragmatic quality (PQ): the inherent usability of a prod- stand their meaning and tried to find it using the control elements. uct that indicates how successful the users can achieve their Based on this finding, we can conclude that the meaning of labels 5 and shades should be made more visible and self-explanatory, so http://www.attrakdiff.de/ Figure 4: Series of gaze fixations of one participant in the first 10 seconds Figure 5: Gaze density of all participants in the first 20 seconds goals with the product. • Hedonic quality - identity (HQ-I): the ability to develop the identity and help the user to establish personal connection with the product. • Hedonic quality - stimulation (HQ-S): the ability to stimu- late the need for further use. • Attractiveness (ATT): the general outward appearance of a Figure 6: Example of word-pairs in the AttrakDiffTM question- product. naire The questionnaire consists of 28 word-pairs (Fig. 6) organized ac- cording to the four qualities. For each pair, the subject has to cast a vote on a seven-value likert scale. 5 out of 10 participants completed the questionnaire online. Fig. 7 shows an overview of the received feedback with respect to the pragmatic (x-axis) and hedonic (y-axis) qualities. As it can be seen from the average value V, the interface was deemed to have a good hedonic quality and an average pragmatic quality. With respect to the hedonic quality, the confidence rectangle shows that the rating is relatively consistent, whereas for the pragmatic quality, it shows a significant deviation. The possible reasons for the significant dis- agreement on the pragmatic quality might be the small number of respondents and the prototypical state of the interface. Fig. 8 shows the mean values for each four qualities influencing the overall attractiveness of IntrospectiveViews. As it can be seen from the figure, the rating of HQ-S is way above average. It shows that the interface has a good ability to develop the stimuli for fur- ther use, which is a very important characteristic of a scrutable user model. HQ-I received a somewhat lower rating than HQ-S. How- ever, it is important to notice, that the identity is strongly affected by the product’s brand and the user’s personal memories and asso- ciations with it. These variables are out of scope of this study. The ratings of the pragmatic quality and attractiveness indicate that the both qualities still have room for improvement. Fig. 9 provides a detailed view on the ratings of each of the four qualities by showing the mean values for the all 28 word pairs. As it can be seen from the diagram, all values are in the positive range. The most interesting information here is the extreme values, which Figure 7: Average rating with respect to the hedonic and prag- show the most problematic characteristics as well as the character- matic qualities istics that have been especially well resolved. Considering the ex- treme positive values, it can be concluded that IntrospectiveViews is a very practical, inventive, innovative, and novel interface. Con- sidering the extreme low values, we see the following problems: First, in the pragmatic quality dimension, two word pairs received relatively low ratings. These are “cumbersome - straightforward” and “confusing - clearly structured”. This is partially caused by such problems as the relatively large number of control elements displayed within the interface, overlapping labels, and poor visibil- ity of important control elements (e.g., controls for adding new in- terests and obtaining detailed information about a specific interest). Another cause of the low ratings in this dimension is the prototypi- cal state of the interface. At the time of evaluation, some functions had not been implemented or were malfunctioning. The rather average rating of the attractiveness is partially caused by the difference in the opinion regarding the color scheme. Half Figure 8: Mean values of the ratings with respect to the four of the participants liked the hot-and-cold metaphor, but the other qualities half found it unpleasant and even dangerous. Another reason is the relatively big number of control elements being displayed, the Figure 9: Mean values of the word pairs small size of the toolbar buttons, and the Java Swing look-and-feel surface. This problem becomes especially serious in large of the interface elements. In the interview, some participants spoke models consisting of several hundreds items. To avoid this out on the Java look rather negatively. problem the interface should provide a mechanism for search- ing items by name. This can be done by means of a search 5. DISCUSSION box, which could be either always displayed in one of the in- In this study we have identified a number of critical usability prob- terface’s corners or invoked by clicking the search button on lems. We believe that researchers working on similar interfaces the toolbar or hitting the conventional [Ctrl + F] on the key- for user model visualization can derive benefits from our findings. board. Having typed the first several characters of the item’s This section provides a summary of the most serious problems and name in the search box, the user would be able to see the outlines possible solutions for resolving them. matching items highlighted on the surface and easily locate the item she is looking for. • Difficulty to understand the main function of the inter- • Excessive amount of control elements. During the inter- face at the first time of use. From the observations we view many subjects complained that their interaction was made during the completion of Task 1 we found that most complicated by the relatively large amount of control ele- participants could not understand the main function and pur- ments displayed in the interface, namely, the buttons on the pose of the interface in isolation by just using it. Since the toolbar and the dialog windows for filtering displayed on the concept of user model is new to the average social network right. To reduce the amount of elements we plan to redesign user, the interfaces representing user models should provide the filtering mechanisms. Instead of selecting the types in the a clear description of their purpose and function. One op- checkbox list of the Types window, the sectors containing the tion to achieve that is to clearly show the user the connection items of a certain types can be minimized/expanded by click- between her user model and the adaptation/personalization ing the type’s label. In a similar way the filtering by interest effects achieved using the model’s content. Also, as it was group can be achieved by clicking the corresponding ring. suggested during the interview, the interface’s elements can E.g., clicking the blue ring in its normal state will minimize be augmented with a short description of their function. Such the ring and hide the items of the group “not interested” and description could be displayed by default for the first time use clicking this ring once again will turn it to the normal state. and invoked later upon request. • Poor visibility of control elements for important func- • Difficulty in finding a specific item. During the task-based tions. Another difficulty that we observed during the com- experiment (Tasks 3-7) and the interview we found that the pletion of tasks was that many subjects did not know about most difficult and time-consuming task while performing a the context menu of items and the background, which contain certain action on a specific item was to locate the item on the such important actions as adding and deleting items, setting privacy settings, viewing semantic relations, etc. One of the to share. Finally, we plan to integrate and evaluate the interface in reasons of being not aware of the context menus is that these a real social network or at least in its prototypical implementation. menus are considered to be a common feature for desktop applications, but is not common on the Web. One option to Acknowledgements let the user know about all the functions available for an item We would like to thank all the participants who took part in the without putting more buttons on the toolbar is to display the usability evaluation. We also want to acknowledge Christoph Ohl list of available actions on the mouse over event. and Alexander Herz for their contribution to this study. • Overlapping items. As we observed during the task-based experiment, a number of participants experienced problems The development of IntrospectiveViews was carried out in the Min- in grasping an item when it was overlapped by other items. erva Context-Adaptive Portals project funded by the IBM Deutsch- This problem was also mentioned by many subjects during land Research & Development GmbH. the interview. In the next version we are going to improve the layout algorithm to minimize the number of overlapped A patent application describing the visualization and interaction items. However, since it is difficult to ensure 100% overlap- concept of IntrospectiveViews has been filed by the IBM Deutsch- ping free layout for visualizing large models on small dis- land Research & Development GmbH. plays, the interface should have a mechanism allowing the user to grasp an item in an overlapped bundle. A possible so- 7. REFERENCES lution to achieve that is repositioning of overlapped items on [1] T. Apted, J. Kay, A. Lum, and J. Uther. Visualisation of the mouse over event. I.e., when the user moves the mouse ontological inferences for user control of personal web over a bundle of overlapped items, they are automatically agents. In E. Banissi, K. Börner, C. Chen, G. Clapworthy, moved in different directions to ensure the appropriate dis- C. Maple, A. Lobben, C. Moore, J. Roberts, A. Ursyn, and tance between them. J. Zhang, editors, Proc. of the 7th Int. Conf. on Information Visualization, pages 306–313, Washington D.C., 2003. IEEE • Different perception of the hot-and-cold metaphor. From Computer Society. the interview we learnt that half of the participants found the [2] F. Bakalov, B. König-Ries, A. Nauerz, and M. Welsch. hot-and-cold metaphor appropriate for the color scheme of Introspectiveviews: An interface for scrutinizing semantic the circular surface, but the other half found it unattractive. user models. In Proc. of the 18th Int. Conf. on User We believe that attractiveness is relatively important charac- Modeling, Adaptation, and Personalization, 2010. teristic of a visualization in the context of a social network. [3] F. Bakalov, B. König-Ries, A. Nauerz, and M. Welsch. Therefore, the interface should allow the user to set the color Scrutinizing user interest models with introspectiveviews. In scheme for the rings that she/he prefers. We plan to add the Adjunct Proc. of the 18th Int. Conf. on User Modeling, traffic light color scheme in addition to the hot-and-cold and Adaptation, and Personalization, 2010. let the user switch between them. [4] F. Bakalov, B. König-Ries, A. Nauerz, and M. Welsch. A hybrid approach to identifying user interests in web portals. 6. CONCLUSIONS AND FUTURE WORK In C. Erfurth, G. Eichler, and V. Schau, editors, Proc. of the In this paper, we have shown how IntrospectiveViews can be used 9th Int. Conf. on Innovative Internet Community Systems, for visualizing and editing semantic user interests models in social volume 148 of LNI, pages 123–134, Bonn, 2009. GI. networks. Through the interface, the user can view the informa- [5] R. Cook and J. Kay. The justified user model: a viewable, tion about her interests that the social network collects and uses for explained user model. In B. Goodman, A. Kobsa, and personalization. In addition, the interface allows the user to cor- D. Litman, editors, Proc. of the 4-th Int. Conf. on User rect and update this information and define how the social network Modeling, pages 145–150, 1994. should use it for personalization effects. [6] V. Dimitrova. STyLE-OLM: Interactive open learner modelling. volume 17, pages 35–78, 2003. Also, we have reported on a usability study of IntrospectiveViews as a visualization of user interests for social networks. We have de- [7] J. Heer and D. Boyd. Vizster: Visualizing online social scribed the results of four evaluation methods, namely, task-based networks. In Proc. of the 9th Intl. Conf. on Information experiment, eye-tracking, interviewing, and a questionnaire. The Visualization, 2005. results show that users deem the interface as a useful and novel [8] T. Hennig. Usability engineering am beispiel einer web tool for social networks and that its way to present and manipulate 3.0-applikation (in german). Master’s thesis, Erfurt user interest models is understandable and easy-to-use. They also University of Applied Sciences, 2010. reveal the existing usability problems of the interface and provide [9] A. Jameson. Adaptive interfaces and agents. In A. Sears and insights on solving them and further improving the interface. J. Jacko, editors, The human-computer interaction handbook: Fundamentals, evolving technologies and In the future work, we plan to solve the identified problems and emerging applications (2nd ed.), pages 433–58. CRC Press, implement the received recommendations in a newer version of In- Boca Raton, FL, 2008. trospectiveViews. We also plan to add new social features in the [10] J. Kay. Scrutable adaptation: Because we can and must. In interface, such as the one allowing the user to compare her model V. Wade, H. Ashman, and B. Smyth, editors, Adaptive with the models of other members or even groups on the social net- Hypermedia and Adaptive Web-Based Systems, volume 4018 work. This can become the means for presenting the community of LNCS, pages 11–19, Heidelberg, 2006. Springer. the user’s personality as an IntrospectiveViews-based visualization [11] J. Kay and A. Lum. Building user models from observations of her interests. It would allow the community members to quickly of users accessing multimedia learning objects. In obtain an overview on the user’s interests with respect to groups, A. Nuernberger and M. Detyniecki, editors, Adaptive topics, locations, and other types of information the user is willing Multimedia Retrieval, pages 36–57, Heidelberg, 2004. Springer. [12] S. Kim. Visualizing Users, User Communities, and Usage Trends in Complex Information Systems Using Implicit Rating Data. PhD thesis, Virginia Tech, 2008. [13] J. Kliger. Model planes and totem poles: Methods for visualizing user models. Master’s thesis, MIT Media Lab, 1995. [14] A. Mabbott and S. Bull. Alternative views on knowledge: Presentation of open learner models. In J. Lester, M. Vicari, and F. Paraguaçu, editors, Proc. of the 7th Int. Conf. on Intelligent Tutoring Systems, volume 3220 of LNCS, pages 689–698, Heidelberg, 2004. Springer. [15] P. Mutton. Inferring and visualizing social networks on Internet Relay Chat. In Proc. of the 8th Intl. Conf. on Information Visualization, 2004. [16] J. Nielsen. Enhancing the explanatory power of usability heuristics. In B. Adelson, S. Dumais, and J. Olson, editors, Proc. of the SIGCHI Conf. on Human Factors in Computing Systems, pages 152–158. ACM, 1994. [17] J. Orwant. Appraising the user of user models: Interface guidelines. In B. Goodman, A. Kobsa, and D. Litman, editors, Proc. of the 4-th Int. Conf. on User Modeling, 1994. [18] B. Shneiderman. The eyes have it: A task by data type taxonomy for information visualizations. In Proc. of the 1996 IEEE Symposium on Visual Languages, pages 336–343, Washington D.C., 1996. IEEE Computer Society. [19] D. Tchuente, M.-F. Canut, N. Jessel, A. Péninou, and A. El Haddadi. Visualizing the evolution of users’ profiles from online social networks. In Proc. of Intl. Conf. on Advances in Social Networks Analysis and Mining (ASONAM), 2010.