Incorporating Tabletop Visual Analytics into the Decision-making Process: A Case Study of Retail Banking Erdem Kaya, Mert Toka, Atilla Bayrak, Burcin Bozkaya, and Selim Balcisoy Fig. 1. A hybrid visual analytics system (a) has been implemented by employing physical visualization and projection of data visual- ization. This approach seems to (b) facilitate a collaborative data analysis space. The projected visualization can be augmented with (c) various visual elements that are bound to data. Abstract—The large-scale use of office tools and statistical analysis applications indicates that they have sufficed well for some of the everyday tasks in our work cycles such as analysis, presentation, reporting, and decision-making. Nevertheless, they were designed in an era when business data was not big and complex enough. The ever growing avalanche of the data that we collect for our businesses compels us to find new means of understanding, sharing, and reporting the underlying ideas, and of making decisions for the future. We implemented a tabletop hybrid visual analytics system comprised of projection and physical visualization with the aim of supporting these tasks better by making the data physically available. We conducted a case study with an analysis team of a nationwide bank and completed a series of observations and interviews during their data analysis and decision-making sessions. Our study revealed that the hybrid visual analytics system approach promotes idea sharing and the contribution of all members of the group during the presentation sessions. This approach also seems to transform the common one-way structure of the communication in the presentation sessions into an alternative structure that encourages everyone to take the floor. Research questions that will be investigated in future work are also discussed. Index Terms—Hybrid visual analytics systems, physical visualization, decision-making, data analysis, case study, retail banking 1 I NTRODUCTION The wide adoption of office and business applications such as data spreadsheets, presentation software, and statistical analysis tools shows that these applications seem to have sufficiently supported the organization and automation of business processes so far. The business data that we have collected has increased so much that various sta- tistical analysis and reporting tools have become an inseparable part of day-to-day operations. Many employees make inferences and re- port their ideas over these tools seamlessly making them means of decision-making processes. The diversity and uniqueness of the sources of our information in- • Erdem Kaya is with Sabanci University, Behavioral Analytics and Visualization Lab. E-mail: erdemkaya@sabanciuniv.edu. • Mert Toka is with Sabanci University, Behavioral Analytics and Visualization Lab. E-mail: merttoka@sabanciuniv.edu. • Atilla Bayrak is with Akbank T.A.S. E-mail: atilla.bayrak@akbank.com. • Burcin Bozkaya is with Sabanci University, Behavioral Analytics and Visualization Lab. E-mail: bbozkaya@sabanciuniv.edu. • Selim Balcisoy is with Sabanci University, Behavioral Analytics and Visualization Lab. E-mail: balcisoy@sabanciuniv.edu. Manuscript received 15 Aug. 2015; accepted 18 Sep. 2015; date of publication 25 Oct. 2015; date of current version 25 Oct. 2015. For information on obtaining reprints of this article, please send e-mail to: tvcg@computer.org. creases our potential for innovation, especially in an age where collab- oration is vital to address business tasks. Consequently, the quality and intensity of social interactions among the members of a group play an important role in creativity and productivity. In the age of Big Data, however, to what extent do our current business applications facili- tate effective idea sharing? Are our channels for sharing information and engaging in decision-making, both socially and technically, able to convey what we extract from Big Data? Can we enhance human- to-human interaction by making use of the techniques available in the Big Data visualization and visual analytics realms? We argue that, in spite of their success in the past, current conven- tional applications fall short in terms of supporting the recall of in- ferences made from the analysis of business data, idea sharing, and decision-making processes. Widely used information visualization and analysis tools seem to fail from time to time in fulfilling idea sharing and decision-making tasks as they are prone to individual in- adequacies. We speculate that the virtual availability of ideas and inferences through the displays of computing devices appears not to assist decision-makers as much as the physical availability of the phe- nomenon in question. We propose that the combination of virtual and physical visualiza- tions can augment data analysis and decision-making by utilizing the best of both worlds. We implemented a tabletop hybrid visual an- alytics system based on projection and physical visualizations with the goal of providing users insights about complex multivariate spa-