Primary care clinicians routinely face ever-expanding, heterogeneous health record corpora while working within chronically short consultation times. To overcome this time-critical data overload challenge, we propose a modular Medical Visual Analytics approach that fuses large language model-driven information extraction with interactive visual decision-support tailored to outpatient workflows. Grounded in an established Visual Analytics process and a widely adopted visualization task model, our approach employs transformer and rule-based NLP pipelines to map unstructured reports onto a unified semantic schema of core clinical data. These entities are presented via a set of visualizations and user interface elements that follow an overview, zoom, and filter paradigm. Our proof-of-concept was evaluated under real-world conditions with a general practitioner using a triangulated approach, demonstrating markedly lower cognitive load, faster retrieval of patient history, and improved guideline adherence under intense time pressure, while also exposing remaining shortcomings. Our work delivers an integrated, primary care-oriented Medical Visual Analytics approach, from conceptual model through working prototype to expert validation. This demonstrates how coupling large language model extraction with task-aligned visualization can measurably enhance decision-making within the strict time constraints of everyday consultations.

Interactive machine learning enhances data pre-processing, feature engineering, and modeling through human interactions. As interactive machine learning becomes more in-corporated into visual analytics, new pipelines, concepts, and task taxonomies rise. A common technique for managing complexity in visual analytic systems is multiple-linked views. By now no comprehensive concept exists for multiple-linked views across all stages of the visual analytic pipeline. Therefore, we introduce a conceptual model that shows the possible interactions between humans and computers at each step of the VA pipeline. Our main contribution is a coordination model for linking multiple views across each layer of the VA pipeline.

Automated visualization generation tools make visualization authoring more accessible to non-programmers and accelerate expert visualization designers in their work. Yet, combining these functions in one system remains a challenge for the research community because it has to maintain a high level of expressiveness and facilitate several tasks for diverse background users while remaining simple and intuitive. Providing a system that handles multiple tasks for heterogeneous user groups could be achieved, on the one hand, through unrestricted user input in the form of natural language. On the other hand, introducing a certain level of abstraction on multiple levels of the system can help integrate more tasks than implementing them separately. In this work, we present the concept of such a system with an LLM-based user query handling and mapping of the extracted components of the user task onto visual design choices with the help of deep learning. Our trained neural network shows promising results, suggesting the generalizability of the proposed approach.

Climate change is one of the greatest challenges of our time and affects all areas of our society. Climate science is a complex field and the majority of the public does not experience the effects of climate change directly but through media. Social media platforms are among the most important communication channels: millions of people exchange information and interact with content on these platforms every day. For this reason, knowledge about the development of climate communication is relevant for various disciplines, but especially for communication sciences. How social media users interact with content on climate change and what content they consume is important for adapting communication strategies. This enables communicators such as journalists to respond appropriately to user interests and align their content. We contribute with a novel approach to portray climate change behavior of social media, transformer-based speech extraction from web video, transformer-based information extraction and trend analysis, and an interactive visual interface showing topic-specific data and trends.

Visual Analytics combines human cognitive processes and abilities with computational models to solve complex problems by using the strengths of computers and humans. Decision-making processes are commonly very complex and involve investigating various factors and indicators. By integrating Visual Analytics into the decision-making process, the human cognitive load can be reduced, and the accuracy of decisions could be higher. This might be why many visual decision support systems exist, and Visual Analytics approaches provide decision support. However, the literature review reveals that there are either structured decision support methods integrating statistical approaches or Visual Analytics methods that are commonly designed for exploration, comparison, and analysis to support decision-making. By investigating the core ideas related to human limitations and programmed decision-making, we propose in this paper a novel Visual Analytics approach to support structured and exploratory approaches by integrating Multi-Criteria Decision Making (MCDM) methods in Visual Analytics. To reduce the common expert judgment in such methods, our approach trains a neural network through the interaction with the system. Our contributions are threefold: providing a systematic view of decision-making processes, discussing the dual paradigms of visual decision support systems, and proposing a Visual Analytics approach that leverages MCDM and neural networks for improved decision support.

The evolving digitization of Germany's medical care provides new opportunities for computer-supported patient treatment. Particularly, resident medical doctors can be empowered to gather comprehensive information about the disease history, medications, and other important indicators quickly and increase the quality of treatment while maintaining the time for treatment. The central element that enables such computer-assisted support in the treatment is electronic health records (EHRs). EHRs facilitate digital access to vast medical data repositories in digital format that were previously confined to analog forms in medical facilities. However, using EHRs in daily medical treatment is time-consuming due to their unstructured textual format. There is a pressing need for analytical tools that extract the most important information from EHRs, provide that information in a quickly comprehensible way, and synthesize the entire patient history for a reliable medical treatment. In this work, we propose an innovative Visual Analytics approach and system specifically designed for medical care. Our approach and the implemented system seamlessly integrate interactive visualizations with fitting pre-trained transformer models to assist medical professionals in consolidating and presenting intricate patient data through comprehensible interactive visual interfaces. Utilizing transformer-based information extraction, it carefully manages the shift from digitized medical documents to quickly accessible patient information in a dynamic and interactive visual interface.

The huge amount of scientific content increases the workload for evaluating state-of-the-art research and the complexity of creating novel and innovative methods and approaches. Although many approaches exist using recommendations in various application domains, the full potential of recommendation systems is not yet fully utilized. Particularly, there are missing approaches that combine interactive visualizations with recommendation systems to enable an analytical investigation of the current state of technology and science. We, therefore, propose in this work a novel Visual Analytics approach that integrates recommendation methods as the model and provides a seamless integration of both interactive visualizations and recommendation systems. We utilize MAE and RMSE metrics and human validation to identify the best approach out of eight approaches that differ in vectorization and similarity algorithms to recommend scientific items. We contribute novel approaches for recommending scientific publications, venues, and projects, based on comparing traditional and deep-learning-based recommendation approaches. Furthermore, we propose a Visual Analytics approach that uses recommendation methods for analytical elaboration. This work shows the potential of integrating recommendation systems into scientific research and identifies potential future directions for improving the proposed model.

Visual Analytics that combines automated methods with information visualization has emerged as a powerful approach to analytical reasoning. The integration of artificial intelligence techniques into Visual Analytics has enhanced its capabilities but also presents challenges related to interpretability, explainability, and decision-making processes. Visual Analytics may use artificial intelligence methods to provide enhanced and more powerful analytical reasoning capabilities. Furthermore, Visual Analytics can be used to interpret black-box artificial intelligence models and provide a visual explanation of those models. In this paper, we provide an overview of the state-of-the-art of artificial intelligence techniques used in Visual Analytics, focusing on both explainable artificial intelligence in Visual Analytics and the human knowledge generation process through Visual Analytics. We review explainable artificial intelligence approaches in Visual Analytics and propose a revised Visual Analytics model for Explainable artificial intelligence based on an existing model. We then conduct a screening review of artificial intelligence methods in Visual Analytics from two time periods to highlight recently used artificial intelligence approaches in Visual Analytics. Based on this review, we propose a revised task model for tasks in Visual Analytics. Our contributions include a state-of-the-art review of explainable artificial intelligence in Visual Analytics, a revised model for creating explainable artificial intelligence through Visual Analytics, a screening review of recent artificial intelligence methods in Visual Analytics, and a revised task model for generic tasks in Visual Analytics.

Knowledge of emerging and declining trends and their potential future course is highly relevant in many application domains, particularly in corporate strategy and foresight. The early awareness of trends allows reacting to market, political, and societal changes and challenges at an appropriate time. In our previous works, we presented approaches for the early identification and analysis of emerging trends. Although our previous approaches are detecting emerging trends appropriately, they lack the ability to predict the potential future course of a trend or technology. We present in this work a novel Visual Analytics approach for forecasting emerging trends that combines interactive visualizations with machine learning techniques and statistical approaches to detect, analyze, and predict trends from textual data. We extend our previous work on analyzing technological trends from text and propose an advanced approach that includes forecasting through hybrid techniques consisting of neural networks and established statistical methods. Our approach offers insights from enormous data sets and the potential future course of trends based on their occurrence in textual data. We contribute with a novel approach for identifying and forecasting trends, a hybrid forecasting method to predict trends from text, and interactive visualization techniques on macro level, micro level, and monitoring topics of interest.

Corporate Foresight is a strategic planning process that helps organizations anticipate and prepare for future trends and developments that may impact their operations. It involves analyzing data, identifying potential scenarios, and creating strategies to address them to ensure long-term success and sustainability. Visual Analytics approaches have been introduced to cover parts of the Corporate Foresight process. These concepts present different approaches to integrate machine learning methods and artificial intelligence with interactive visualizations to solve tasks such as identifying emerging trends. A holistic concept for synthesizing Visual Analytics with Corporate Foresight does not exist yet. We propose in this work a holistic Visual Analytics approach that covers the main aspects of Corporate Foresight by including strategic management and considers different organizational forms. Our model goes beyond the state-of-the-art by providing, besides foresight also, hindsight and insight. Our main contributions are the revised Visual Analytics model and its proof of concept through implementation as a web-based system with real data.

Studies have shown that collaboration in scientific fields is rising and considered enormously important. However, collaboration has proved to be challenging for various reasons, among others, the requirements for human-machine workflows. The importance of scientific collaboration lies in the complexity of the challenges that are faced today. The more complex the challenge, the more scientists should work together. The current form of collaboration in the scientific community is not as intelligent as it should be. Scientists have to multitask with various applications, often losing cognitive focus. Collaboration itself is very nearsighted as it is usually conducted not solely based on expertise but instead on social or local networks. We introduce a single-source visual collaboration approach based on learning methods in this work. We use machine learning and natural language processing approaches to improve the traditional research and development process and create a system that facilitates and encourages collaboration based on expertise, enhancing the research collaboration process in many ways. Our approach combines collaborative Visual Analytics with enhanced collaboration techniques to support researchers from different disciplines.

Systematic reviews play an essential role in various disciplines. Particularly, in biomedical sciences, systematic reviews according to a predefined schema and protocol are how related literature is analyzed. Although a protocol-based systematic review is replicable and provides the required information to reproduce each step and refine them, such a systematic review is time-consuming and may get complex. To face this challenge, automatic methods can be applied that support researchers in their systematic analysis process. The combination of artificial intelligence for automatic information extraction from scientific literature with interactive visualizations as a Visual Analytics system can lead to sophisticated analysis and protocoling of the review process. We introduce in this paper a novel Visual Analytics approach and system that enables researchers to visually search and explore scientific publications and generate a protocol based on the PRISMA protocol and the PRISMA statement.

Abstract The awareness of emerging trends is essential for strategic decision making because technological trends can affect a firm’s competitiveness and market position. The rise of artificial intelligence methods allows gathering new insights and may support these decision-making processes. However, it is essential to keep the human in the loop of these complex analytical tasks, which, often lack an appropriate interaction design. Including special interactive designs for technology and innovation management is therefore essential for successfully analyzing emerging trends and using this information for strategic decision making. A combination of information visualization, trend mining and interaction design can support human users to explore, detect, and identify such trends. This paper enhances and extends a previously published first approach for integrating, enriching, mining, analyzing, identifying, and visualizing emerging trends for technology and innovation management. We introduce a novel interaction design by investigating the main ideas from technology and innovation management and enable a more appropriate interaction approach for technology foresight and innovation detection.

Visual Analytics enables solving complex analytical tasks by coupling interactive visualizations and machine learning approaches. Besides the analytical reasoning enabled through Visual Analytics, the exploration of data plays an essential role. The exploration process can be supported through similarity-based approaches that enable finding similar data to those annotated in the context of visual exploration. We propose in this paper a process of annotation in the context of exploration that leads to labeled vectors-of-interest and enables finding similar publications based on interest vectors. The generation and labeling of the interest vectors are performed automatically by the Visual Analytics system and lead to finding similar papers and categorizing the annotated papers. With this approach, we provide a categorized similarity search based on an automatically labeled interest matrix in Visual Analytics.

The integrated cyber-physical systems in Smart Manufacturing generate continuously vast amount of data. These complex data are difficult to assess and gather knowledge about the data. Tasks like fault detection and diagnosis are therewith difficult to solve. Visual Analytics mitigates complexity through the combined use of algorithms and visualization methods that allow to perceive information in a more accurate way. Thereby, reasoning relies more and more on the given situation within a smart manufacturing environment, namely the context. Current general Visual Analytics approaches only provide a vague definition of context. We introduce in this paper a model that specifies the context in Visual Analytics for Smart Manufacturing. Additionally, our model bridges the latest advances in research on Smart Manufacturing and Visual Analytics. We combine and summarize methodologies, algorithms and specifications of both vital research fields with our previous findings and fuse them together. As a result, we propose our novel industry 4.0-ready Visual Analytics model for context-aware diagnosis in Smart Manufacturing.

Scientific publications are an essential resource for detecting emerging trends and innovations in a very early stage, by far earlier than patents may allow. Thereby Visual Analytics systems enable a deep analysis by applying commonly unsupervised machine learning methods and investigating a mass amount of data. A main question from the Visual Analytics viewpoint in this context is, do abstracts of scientific publications provide a similar analysis capability compared to their corresponding full-texts? This would allow to extract a mass amount of text documents in a much faster manner. We compare in this paper the topic extraction methods LSI and LDA by using full text articles and their corresponding abstracts to obtain which method and which data are better suited for a Visual Analytics system for Technology and Corporate Foresight. Based on a easy replicable natural language processing approach, we further investigate the impact of lemmatization for LDA and LSI. The comparison will be performed qualitative and quantitative to gather both, the human perception in visual systems and coherence values. Based on an application scenario a visual trend analytics system illustrates the outcomes.

The awareness of emerging technologies is essential for strategic decision making in enterprises. Emerging and decreasing technological trends could lead to strengthening the competitiveness and market positioning. The exploration, detection and identification of such trends can be essentially supported through information visualization, trend mining and in particular through the combination of those. Commonly, trends appear first in science and scientific documents. However, those documents do not provide sufficient information for analyzing and identifying emerging trends. It is necessary to enrich data, extract information from the integrated data, measure the gradient of trends over time and provide effective interactive visualizations. We introduce in this paper an approach for integrating, enriching, mining, analyzing, identifying and visualizing emerging trends from scientific documents. Our approach enhances the state of the art in visual trend analytics by investigating the entire analysis process and providing an approach for enabling human to explore undetected potentially emerging trends.

Nowadays a wide range of input devices are available to users of technical systems. Especially modern alternative interaction devices, which are known from game consoles etc., provide a more natural way of interaction. In parallel to that the research on visualization of large amount of data advances very quickly. This research was also influenced by the semantic web and the idea of storing data in a structured and linked form. The semantically annotated data gains more and more importance in information acquisition processes. Especially the Linked Open Data (LOD) format already experienced a huge growth. However, the user-interfaces of web-applications mostly do not reflect the added value of semantics data. This paper describes the conceptual design and implementation of an Interactive Exploration System that offers a user-centered graphical environment of web-based knowledge repositories, to support and optimize explorative learning, and the integration of a taxonomy-based approach to enable the use of more natural interaction metaphors, as they are possible with modern devices like Wii Mote or Microsoft Kinect. Therefore we introduce a different classification for interaction devices, and current approaches for supporting the added values in semantics visualizations. Furthermore, we describe the concept of our IES, including a strategy to organize and structure today's existing input devices, and a semantics exploration system driven by user-experience. We conclude the paper with a description of the implementation of the IES and an application scenario.