Multi-tag-based search is quite popular on collaborative websites and sharing-online-content systems. For this kind of search results, the challenge is how to compare grouped-tag values of tag collections on heterogeneous alternatives. This paper introduces a new visualization approach named Qstack for dealing with the challenge. Qstack purpose is to help users to visually rank multi-tags based on grouped-score combination within and across the categorized alternatives. The methodology applying interactive stacked bars, dynamic queries and adaptive focus+context techniques enables users to easily create and adjust grouped-tag rankings of a large number of heterogeneous alternatives. A case study on Flickr photo award allocation will be presented for Qstack demonstration. We conducted a qualitative study for evaluating Qstack effectiveness, and the result indicates that our approach is useful for multi-tag rankings.

No abstract

Parallel coordinates and scatter plots are two well-known visualization techniques for multidimensional data analytics and often employed cooperatively for flexibility increase in exploration of such data. Existing approaches approximately consider qualitative issues and single attribute comparison, which might face statistic challenges in case of quantitative requirement. This paper introduces a new quantitative approach for visual enhancement of parallel coordinates and scatter plots in term of multiple attribute comparison. The method is based on the visual integration of interactive stacked bars and visual queries on parallel axes and scatter charts. The parallel coordinates play the role of a context view while the scatter charts are for focus details. Using the technique, users could not only quantitatively analyze multivariate data, but also flexibly compare multiple target attributes. Moreover, further investigation is enabled for deep understanding of desired information. The characteristics and usefulness of our approach are demonstrated via a case study with two typical use cases.

No abstract

No abstract