Scientific Visualization, 2018, volume 10, number 4, pages 82 - 99, DOI: 10.26583/sv.10.4.07
Visual Analytics Tools for Systematic Exploration of Multi-Parameter Data of Social Web-Based Service Users
Authors: K.V. Ryabinin1*, K.I. Belousov2*, S.I. Chuprina3*, S.A. Shchebetenko4+*, S.S. Permyakov5*
* Perm State University, Perm, Russia
+ National Research University Higher School of Economics, Moscow, Russia
1 ORCID: 0000-0002-8353-7641, kostya.ryabinin@gmail.com
2 ORCID: 0000-0003-4447-1288, belousovki@gmail.com
3 ORCID: 0000-0002-2103-3771, chuprinas@inbox.ru
4 ORCID: 0000-0001-5790-9731, shebetenko@rambler.ru
5 ORCID: 0000-0002-8486-779X, rewmad@gmail.com
Abstract
The paper is devoted to the usage of the visual analytics methods and means for systematiñ exploration of the results of a multi-parameter data of social Web-based service users. These data include language characteristics of the users’ comments and posts obtained from the social services they use, as well as psychological and social characteristics obtained from their profiles and from the results of surveys they fulfilled. Suggested visual analytics tools enable to present the correlations between different users’ characteristics in an observable form and to help proposing and testing hypotheses without repeating the initial data processing experiment, using just visual analytics tools to produce new results. This in turn enables to uncover and study the regularities in the input data. Semograph linguistic analysis software is suggested as a tool to collect and preprocess the data. SciVi ontology-driven multiplatform adaptive scientific visualization system is proposed to be a visual analytics tool.
Because the input data have a lot of interconnections, the described visual analytics tools are based on a graph data representation model. Circular graph with adjustable hierarchical ring scale and free structured graph are supported within SciVi to ensure advanced visual analytics. The paper presents the main interactive features and implementation details of suggested tools. In particular, different filtering mechanisms for nodes and arcs are presented, as well as the means to navigate through different input data slices.
The implemented visual analytics tools are tested by solving the real world problems related to the results of psychological surveys. The survey was conducted among the users of VKontakte social network. The dependencies between their psychological characteristics and verbal behavior are discovered.
Keywords: visual analytics, ontology engineering, graphs, lingual parameters, psychological characteristics, BFI, social networks users parameters.
