FIGURE 1 - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
Source publication
The need for efficient and effective data exploration has resulted in several solutions that automatically recommend interesting visualizations. The main idea underlying those solutions is to automatically generate all possible views of data, and recommend the top-k interesting views. However, those solutions assume that the analyst is able to form...
Contexts in source publication
Context 1
... mentioned in Section IV, to enable the refinement of continuous dimensions, those dimensions are discretized by a user specified parameter γ. Figures 12 and 13 show the impact of this discretization factor, or grid resolution, on the cost and overall utility value achieved by all schemes. Note that a large value of γ represents a sparse grid with wider cells, which means there are fewer refined queries to explore. ...
Context 2
... refined queries generally implies increase in cost but it should also be able to improve the overall utility of the top-k views. Figure 12 clearly shows that as the grid becomes dense the cost of the Linear scheme increase as it has to search through a larger number of refined queries for top-k views. At the same time, the reduction in cost by the QuRVe schemes are more prominent for dense grids (i.e., low γ). ...
Context 3
... is because the number of possible refined queries increases and it gives the QuRVe schemes more opportunities to prune unnecessary views. Figure 12b shows the same results of Figure 12a by using a log scale on the y-axis to further underscore the performance gains provided by QuRVe. For instance, the figure clearly shows that at γ = 0.0625, the cost of pQuRVe is more than 10 folds less than that for Linear. ...
Context 4
... is because the number of possible refined queries increases and it gives the QuRVe schemes more opportunities to prune unnecessary views. Figure 12b shows the same results of Figure 12a by using a log scale on the y-axis to further underscore the performance gains provided by QuRVe. For instance, the figure clearly shows that at γ = 0.0625, the cost of pQuRVe is more than 10 folds less than that for Linear. ...
Citations
... That challenge motivated multiple research efforts that focused on automatic recommendation for data exploration. That is, recommender systems dedicated to providing the user with suggestions for specific, high-utility visualizations (e.g., [17][18][19][20][21][22][23][24][25][26][27][28]). Such systems are data-driven (also known as discovery-driven) systems, which use heuristic notions of "interestingness" and employ them in the recommendation. ...
... However, current visualization recommendation systems (e.g., [18,26,27]) assume that the analyst is able to formulate a well-defined query that selects a subset of data, which leads to insightful visualizations being recommended (i.e., visualizations with a high utility score). That is, they are limited to only recommending interesting visualizations based on a precise exploratory query for which the analyst provides all the necessary query filters. ...
Analyzing crowdfunding data has been the focus of many research efforts, where analysts typically explore this data to identify the main factors and characteristics of the lending process as well as to discover unique patterns and anomalies in loan distributions. However, the manual exploration and visualization of such data is clearly an ad hoc, time-consuming, and labor-intensive process. Hence, in this work, we propose LoanVis, which is an automated solution for discovering and recommending those valuable and insightful visualizations. LoanVis is a data-driven system that utilizes objective metrics to quantify the “interestingness” of a visualization and employs such metrics in the recommendation process. We demonstrate the effectiveness of LoanVis in analyzing and exploring different aspects of the Kiva crowdfunding dataset.
... Data-View Recommendation: Our work in this paper is related to the topic of automati-cally identifying and recommending interesting data views to facilitate data exploration, with the intent of maximizing insights from, and/or the value of, the underlying data to the users, see, e.g., [18][19][20][21][22][23][24][25][26]. Existing approaches are often based on data cubes [6,7] and work with a variety of granularity levels of the views to be identified. ...
... Existing works in this area also vary in their definitions of scoring functions that de-termine the level of interestingness of candidate views. Options that have been considered include the level of deviation of a view from reference views [20], as well as the outputs of statistical analyses [23,24] or of techniques based on machine learning [25], among oth-ers [22,21]. The scoring functions tend to be completely or partially predefined in these works, whether based on the data-analytics experience of the authors, on the results of user studies, or on feedback from real-word data analysts [26]. ...
Locating unusual temporal trends in data cubes is a recurrent task in a variety of application domains. We consider a version of this problem in which one looks for the data dimensions that are best correlated with the given unusual temporal trends. Our goal is to make such data-cube navigation in search of unusual temporal trends both effective and efficient. Challenges in achieving this goal arise from the rarity of the trends to be located, as well as from the combinatorics involved in locating in data cubes nodes with unusual trends. We show that exhaustive solutions are worst-case intractable, and introduce tractable heuristic algorithms that enable effective and efficient data-cube navigation in a particular manner that we call trend surfing. We report the results of testing the proposed algorithms on three real-life data sets; these results showcase the effectiveness and efficiency of the algorithms against the exhaustive baseline.
... Information discovery is an emerging search and exploration paradigm [3], [45]. It is a challenging task to pose an exploratory search leading towards discovery [46]. The previous studies measured the effectiveness of advanced proposed search engines by comparing them with a generalization of traditional search engines [45]. ...
The recent advancements in the web allow users to generate multimedia content, resulting in multimedia information proliferation. Existing search engines provide access to multimedia content via a disjoint assembly of media-specific results called verticals. However, this decentralized assembly of media contents requires manual aggregation and synthesizing efforts at the user’s end, hindering the information exploration process and subsequently may cause cognitive overload, hence, demanding innovative tools to discover multimedia content. The researchers have devised numerous state-of-the-art approaches; however, analysis to confirm the efficacy has little emphasis. This study investigates users’ complex multimedia information-seeking behavior over state-of-the-art web search systems to unveil the user’s information-seeking issues. Our research employs between-subjects study and post hoc analysis strategies to analyze participants’ information-seeking characteristics. The study design adopted statistical hypothesis testing to consolidate previous user behavioral studies, confirm existing strategies, and present recommended practices for future general-purpose web search engines. The participants were assigned Google and an advanced discovery search system using the same multimedia dataset to ensure the obtained results’ credibility. The primary behavioral parameters include search efforts, multimedia content exploration, search user interface (SUI), information management and presentation, and user cognition. This study uncovers several inadequacies of the search engines in meeting users’ complex discovery needs, including 29.6% less user engagement, 43% system and searching dissatisfaction, and 32% less knowledge acquisition with 63.9% increased clicking effort on traditional search engines. The results confirmed previous user studies and suggest novel research recommendations statistically significant in multimedia information exploration-related endeavors.
