The vision of Information and Software Visualization (ISOVIS) is to attack the big data challenge by a combination of human-centered data analysis and interactive visualization for decision making. These research topics are highly relevant for academia and economy, as both science and industry make increasing use of data-intensive technologies.
Human-centered visualization deals with the development of interactive visualization techniques in consideration of user- and task-related information to explore and analyze complex data sets efficiently. Sensor data measured during the usage of a visualization (from brain-computer interfaces, eye trackers, etc.) may also be involved. This approach combines aspects of different research areas, such as information and scientific visualization, human-computer interaction, information design and cognition, but also the particular application field. From all subfields of visualization, we mainly focus on information visualization (InfoVis) which centers on the visualization of abstract data, e.g., hierarchical, networked, or symbolic information sources.
In contrast to visualization, data mining (DM) or machine learning (ML) are traditionally more computer-centered. But to address the big data challenge, we have to use the advantages of both approaches synergistically, which is the main feature of visual analytics (VA). Then, the analyst can focus his/her perceptual and cognitive capabilities on the analytical processes while using advanced computational methods to support and enhance the discovery process. The design and implementation of visual analytics tools is one of the most promising approaches to cope with the ever increasing amount of data produced every day and allows new insights and beneficial discoveries.
Our research expertise and interests cover the following areas:
- Text visualization and visual text analytics
- Explainable AI/ML using visualization
- Network visualization and visual network analytics
- Multidimensional data visualization
- Foundations of visualization
- Software visualization
- Human-computer interaction
In projects, we usually combine the expertise of people coming from several fields: from interactive visualization and visual analytics to machine learning and domain experts such as linguists or social scientists. We are always open to collaborate with external partners who provide challenging data sets together with interesting research questions.
Learn more at ISOVIS' web site.
Connection to Linnaeus University Centre for Data Intensive Sciences and Applications (DISA)
The Linnaeus University Centre for Data Intensive Sciences and Applications (DISA) focuses its efforts on open questions in collection, analysis and utilization of large data sets. With its core in computer science, it takes a multidisciplinary approach and collaborates with researchers from all faculties at the university.
Andreas Kerren, chair of ISOVIS, was co-applicant of DISA and is member of the DISA coordination board. Visualization and visual analytics are key technologies for (big) data analysis and therefore naturally connected to DISA and its main goals. The broad expertise of the ISOVIS members within several subfields of information visualization and visual analytics reflects the diversity of data sets and analysis problems that analysts/researchers from all faculties of the university are facing in their daily work. ISOVIS can provide answers to many of such analytical challenges.
As applied machine learning approaches are more and more common in DISA, we also contribute by making the black boxes of the analysis methods and resulting models transparent for the analyst. For this, and for providing reliable trust into the models, we develop foundational visual analytics principles, techniques and tools for analyzing both data and machine learning models.
Connection to Smarter Systems
Smarter Systems is a complete knowledge environment focusing on the systems and challenges of the future. It addresses challenges together with its partners through education and research initiatives in collaboration. Developers of complex systems need to be able to understand the system structure and behavior at any point in time, even when the system has evolved based on the learned knowledge.
Here, visual analytics is crucial for improving the understanding of the complex, dynamic and evolving data and processes that occur at both levels: when engineering smarter systems as well as during their operation. In consequence, we as humans are then able to understand their structure and behavior anytime, even when they have evolved based on the learned knowledge.
In recent years, as a result of developments in sensor and communication technologies, the amount of data collected by complex systems has increased exponentially. Engineers, operators and other stakeholders use these data sets to monitor the processes that systems are situated in or within the system itself. However, standard tools that visualize (or analyze) the data do not provide the required capabilities and analytical features: new means to make effective use of this data are needed.
There are three grand challenges in this context that are addressed by ISOVIS:
Dealing with the four Vs of big data analysis (volume, variety, veracity, velocity) by using visual analysisSmarter systems set specific challenges to the analysis of data, such as uncertainty in the data or the sheer size and complexity of real-world data. Those issues in combination with temporal (streaming) aspects put high demands on a visual analysis (VA) solution.
Tackling the first grand challenge will contribute to achieving key objective 2 for the knowledge environment Smarter Systems: To devise engineering processes for perpetual adaptation and evolution. In particular, advanced VA techniques will contribute to the design and application of engineering processes for perpetual adaptation and evolution.
Transfer of analytics methods, processes and tools to smarter systemsMost current VA tools were solely developed and aligned to analytics tasks for specific applications. The reuse of VA components, and the transfer from other application domains to Smarter Systems, is highly desirable but particularly challenging. Tackling our second grand challenge will contribute to achieving key objective 2 by transferring analytics methods, processes and tools of established domains to Smarter Systems, supporting the engineering processes.
- Explainability of machine learning models by using visual analytics
Opening the black box of computational models, such as supervised or unsupervised machine learning, is a recent challenge in VA. This area is also called explainable AI in general, and visualization tools for machine learning are able to increase the trust into such methods. Tackling the third grand challenge will contribute to both assuring product capabilities in uncertain conditions (key objective 1) and devising smart engineering processes (key objective 2). This is done through the development of proper methods and techniques to better understand the systems and to support their operators in concrete real-world settings.
Connection to Education
ISOVIS is responsible for the following courses:
- 1DV437, Introduction to Game Programming, 7.5 credits
- 1DV512, Operating Systems, 7.5 credits
- 1DV513/2DV513, Database Theory, 7.5 credits
- 1DV800, Computer Graphics, 7.5 credits
- 1ME326, Web Technology 6, 7.5 credits
- 2DV505, Current Topics within Computer Science, 7.5 credits
- 2ME301, Scientific Methods in Media Technology, 7.5 credits
- 4DV504, Selected Topics in Computer Science, 5 credits
- 4DV507, Code Transformation and Interpretation, 5 credits
- 4DV510, Data Mining, 5 credits
- 4DV805, Information Visualization, 5 credits
- 4DV806, Advanced Information Visualization and Applications, 5 credits
- 4DV807, Project in Visualization and Data Analysis, 10 credits
- 4DV808, Computational and Visual Text Analysis, 5 credits
- 4DV809, Computational and Visual Network Analysis, 5 credits
- 4ME302, Foundations of Computational Media, 7.5 credits
- 4ME305, Web and Mobile Development, 7.5 credits
Doctoral project: Multivariate network embedding for visual analytics This doctoral project aims to explore the use of embedding technologies for visual and computational analytics on large…
Doctoral project: Supporting Synchronous Asymmetric Data Exploration using Virtual Reality within the context of Collaborative Immersive Analytics Immersive technologies, such as virtual reality,…
Doctoral project: Visual analytics for explainable and trustworthy machine learning This doctoral project aims to develop foundational principles, techniques, and tools for visual analytics, for…
Doctoral project: Visual learning analytics to support teachers’ pedagogical work This doctoral project features a collaboration with a number of EdTech companies that develop digital learning…
Project: InfraVis – National Research Infrastructure for Visualisation of Data InfraVis will create a common gateway to Swedish visualization resources and enhance the competence of researchers in…
Project: Visualization and Exploration Flexiboard for Timber Buildings (TimberVis) Within the collaboration between InfraVis, the national research infrastructure for data visualization, and the…
Project: StaViCTA Stancetaking is an important factor for social interaction in human communication. This interdisciplinary project will identify how we express stance on the Internet – to create a…
Seed project: Data intensive analysis for identification and prediction of risk medications The main objective for this seed project within Linnaeus University Centre for Data Intensive Sciences and…