In today’s digitized world, an overwhelming amount of still growing resources such as websites, images, texts, figures, and videos are generated. The resulting Big Data Problem does not only consist of the handling of this immense volume of data. Moreover, data needs to be processed, cleaned, and presented in a user-friendly, graphical way. Industrial applications can leverage the potential of this variety of information, thus offering more reliable, diverse, and useful services to customers.

Since dealing with Big Data is inherently complex, many disciplines in computer science and mathematics need to work together. Visual Analytics provides the necessary automated analysis techniques in combination with interactive visualizations in order to effectively understand large and complex datasets (Keim2008). Bertini defines the foundations of Visual Analytics as information visualization, data mining, and interactive machine learning (Bertini2009). While information visualization targets the visual representation of large-scale data collections, data mining aims at extracting hidden patterns and models from data in an automatic or semi-automatic way. Oliveira stresses the fact that data mining can be part of a more general process of extracting high-level, potential useful knowledge, from low-level data, known as Knowledge Discovery in Databases (KDD). Visual techniques are needed for Model Visualization and Exploratory Data Analysis (EDA) to make the discovered knowledge understandable and interpretable by humans (Oliveira 2003). Finally, interactive machine learning advocates the integration of human and machine capabilities to build better computational models (Bertini2009).

This workshop addresses interfaces for three different user groups that are concerned with Big Data problems in industrial contexts. First, data scientists need to select, prepare and tune the algorithms needed to extract information from large data sets. This information includes clusters, classifications, and meta information about the data structures. To this end, suitable visualizations are needed to assess the way algorithms and their parameters generate the results. Second, the often needed involvement of human intervention by data workers is addressed to improve the underlying algorithms. These data workers need specialized views to inspect and compare concrete data items in a quick and effective manner. Third, end users are often overwhelmed by exploring and searching within large data sets. Today, they often have to deal with result lists produced by recommender systems working on the Big Data clusters without knowledge about the way recommendations are generated while in the future, feature-rich, adaptive, and intuitive interfaces will be expected.

Topics of Interest

Based on these three user groups and the described issues, the workshop is intended to provide a forum for discussing research questions and solutions in industrial applications. We invite position papers between 4 and 6 pages in length, focused on the following topics:

  • Visual Interfaces for Big Data Landscapes, e.g.
    • Interaction and Visualization Technologies for Exploring Big Data Repositories
    • Visual Support for Transparency in Big Data and Recommender Systems
    • Visual Representation of Linked and Big Data
    • Interfaces and Interaction Patterns for Visual Analytics
  • Visual Search Interfaces, e.g.
    • Interfaces for Representing Natural Language Queries
    • Explorative Search in Large Data Sets
  • Visualization Tools for enabling Human-in-the-loop Processes, e.g.
    • Visualization Tools for Semi-supervised Machine Learning or Cluster Analysis
    • Visualization and Analysis of Multi-dimensional Scaling Algorithms
    • Visual Tools for Crowd working
  • Interactive Processes for Data Integration and Cleansing, e.g.
    • Interactive Record Linkage
    • Facilitating Record Linkage with Visual Approaches
  • Visual and Interactive Support in Big Data and Recommender Systems, e.g., providing Transparency or Controllability
  • Best Practices for Representing Heterogeneous Data Sources
  • Integration of Implicit and Explicit User Feedback
  • Industrial Use Cases in the Mentioned Fields

References

(Bertini 2009)  Enrico Bertini and Denis Lalanne. 2009. Surveying the Complementary Role of Automatic Data Analysis and Visualization in Knowledge Discovery. In Proceedings of the ACM SIGKDD Workshop on Visual Analytics and Knowledge Discovery: Integrating Automated Analysis with Interactive Exploration (VAKD ’09) . ACM, New York, NY, USA, 12–20. https://doi.org/10.1145/1562849.1562851

(Keim 2008)   Daniel Keim, Gennady Andrienko, Jean-Daniel Fekete, Carsten Görg, Jörn Kohlhammer, and Guy Melançon. 2008. Visual Analytics: Definition, Process, and Challenges. Springer Berlin Heidelberg, Berlin, Heidelberg, 154–175. https://doi.org/10.1007/978-3-540-70956-5_7

(Oliveria 2003)  M. C. Ferreira de Oliveira and H. Levkowitz. 2003. From Visual Data Exploration to Visual Data Mining: A Survey. IEEE Transactions on Visualization and Computer Graphics 9, 3 (July 2003), 378–394. https://doi.org/10.1109/TVCG.2003.120744