Comparing a Grid-based vs. List-based Approach for Faceted Search of Social Media Data on Mobile Devices

In this paper, we compare two approaches for exploring large, hierarchical data spaces of social media data on mobile devices using facets. While the first approach arranges the facets in a 3x3 grid, the second approach makes use of a scrollable list of facets for exploring the data. We have conducted a between-group experiment of the two approaches with 24 subjects (20 male, 4 female) executing the same set of tasks of typical mobile users’ information needs. The results show that the grid-based approach requires significantly more clicks, but subjects need less time for completing the tasks. Furthermore, it shows that the additional clicks do not hamper the subjects’ satisfaction. Thus, the results suggest that the grid-based approach is a better choice for faceted search on touchscreen mobile devices. To the best of our knowledge, such a summative evaluation of different approaches for faceted search on mobile devices has not been done so far. INTRODUCTION The Web 2.0 has brought an explosion of social media data to mobile devices. Exploring this social media data while being on-the-go is an important task to, e.g., find answers to trivia questions like retrieving information about a building or landmark, finding locations like sights or restaurants, or events (cf. [15]). Aggregating social media data from various Web 2.0 sources like Wikipedia1, event directories such as Upcoming2 and Eventful3, places and restaurants from Qype4, and images from Flickr5 results in a large, hierarchical data space. Exploring such a large, hierarchical data space is a difficult task, which becomes even more challenging when limitations of mobile devices like smaller display size and limited interaction possibilities have to be taken into account. Thus, an appropriate approach is needed to find the right piece of information in the social media data space. Providing an application that supports the mobile users in exploring the social media data space is an interesting research question, as the social media data can provide answers to many typical mobile users information needs [15]. To alleviate the problem of finding the right piece of information, the approach of faceted search has been developed in http://wikipedia.org/ http://upcoming.yahoo.com/ http://eventful.com/ http://qype.com/ http://flickr.com/ the past. Initially designed for desktop computers [18, 5, 2], the idea of faceted search has been transferred to mobile devices [12, 11, 4]. A facet can be understood as a dimension in the data space and allows for better exploring the data by limiting the amount of data instances shown to the user. Facets can be arranged hierarchically, i.e., they can be organized in a tree-like manner. Thus, a facet can have one or multiple subfacets. A sub-facet can be associated to more than one facet, resulting in a poly-hierarchical tree. Finally, a facet that does not have any sub-facet is called a leaf facet (cf. [14]). The existing applications for faceted search on mobile devices differ in their approach how to arrange the facets on the screen and using these facets for exploring the data space. The FaThumb system [11] developed by Microsoft makes use of a 3x3 grid for arranging the facets. On the contrary, one finds applications pursuing a list-based approach for faceted search on mobile devices [12, 4]. While the grid-based approach by design starts with a full result list of resources, the initial result list of the list-based approach is empty. In addition, the amount of space used for displaying the facets differs between the approaches. In this paper, we compare both approaches for faceted search on touchscreen mobile devices with respect to their usability, i.e., effectiveness, efficiency, and satisfaction. To the best of our knowledge, such a comparison has not been conducted so far. Thus, no clear recommendation for mobile faceted search has emerged yet [9]. We have developed MFacets following the grid-based approach and MobileFacets providing the listbased approach for faceted search. Goal of this comparison is to investigate the differences of both approaches regarding their usability. To this end, we formulate the null hypothesis saying that there is no significant difference between the grid-based approach for faceted search and the list-based approach for faceted search regarding effectiveness, efficiency, and satisfaction. In order to investigate this hypothesis, we have conducted a summative evaluation where two independent groups of subjects have conducted the same set of tasks but each under different conditions. One group has used the grid-based MFacets and the other group has worked with the list-based MobileFacets. In total, 24 subjects (20 male, 4 female) have participated in this evaluation. Our results show that the grid-based approach requires significantly more clicks (ttest/Mann-Whitney U, P ≤ 5%). However, the tasks are executed faster than using the list-based approach. Thus, one 3 Comparing a Grid-based vs. List-based Approach for Faceted Search of Social Media Data on Mobile Devices, Fachbereich Informatik 1/2012