Hybrid Open Access
Filtern
Dokumenttyp
- Wissenschaftlicher Artikel (36)
- Konferenzveröffentlichung (1)
- Preprint (1)
Gehört zur Bibliographie
- ja (38)
Schlagworte
- Visualisierung (8)
- Journalismus (4)
- Partizipation (4)
- Transdisziplinarität (4)
- Kind (3)
- Kleinkind (3)
- Künstliche Intelligenz (3)
- Nachhaltigkeit (3)
- Rechtswissenschaft (3)
- Bildung (2)
Institut
- FB1 Sozial- und Bildungswissenschaften (16)
- Inst. für angewandte Forschung Urbane Zukunft (IaF) (15)
- FB4 Design (12)
- FB5 Informationswissenschaften (4)
- Publikationen des FB Informationswissenschaften (4)
- Urban Complexity Lab (UCLAB) (3)
- FB3 Bauingenieurwesen (2)
- Hochschulleitung (1)
- Urbane Zukunft (1)
The affective connotations of words are central to meaning and important predictors of many social processes. As such, understanding the degree to which commercially-available generative language models (LLMs) replicate human judgements of affective connotations may help better understand human-model interactions. LLMs may also serve as useful tools for researchers seeking affective meaning estimates. We test the ability of three LLMs – GPT-4o, Mistral Large, and Llama 3.1 – to estimate human affective connotation ratings of words representing social identities, behaviours, modifiers, and settings in three language cultures: English (US), French (France), and German (Germany). We find that LLM ratings of terms correlate strongly with human ratings. However, their ratings tend to be overly extreme and patterns of correlations between meaning dimensions only loosely approximate those of human ratings. Consistent with previous findings of English-language and American biases in LLMs, we find that LLMs tend to perform better on English terms, though this pattern varies somewhat by meaning dimension and the type of term in question. We explore how LLMs might contribute to scholarship on affective connotations – by acting as tools for measurement – and how scholarship on affective connotations might contribute to generative language models – by guiding exploration of model biases.
The persistently lower participation of women in entrepreneurial activities constitutes a significant economic and societal problem. This study investigates the entrepreneurial intentions (EI) of university students and graduates from the University of Applied Sciences Potsdam from a gender-specific perspective, with the central aim of identifying the reasons for the continuing differences in EIs between men and women. The theoretical framework is based on the Theory of Planned Behavior (TPB), which was extended in this study by the construct of entrepreneurial self-identity to gain a more comprehensive understanding of EI. Using structural equation modeling, an analysis was conducted for the entire cohort as well as stratified by gender, examining how attitude, social norms, perceived behavioral control, and entrepreneurial self-identity influence EI. The results of the online survey, in which 414 students and graduates participated between June 2nd and July 14th, 2023, show that men exhibit significantly higher EIs than women. Furthermore, it was demonstrated that risk-taking propensity contributes to these gender differences. Moreover, perceived behavioral control and parental role models proved to be crucial determinants for the EIs of women, while these factors did not significantly relate to the intentions of men. The analysis underscores the complexity of gender differences in EIs and implies that considering gender-specific factors in targeted support measures for universities can help reduce gender differences and thus better leverage the entrepreneurial potential of women.
This case study presents an innovative approach for explaining wildfire susceptibility through a web-based Geospatial eXplainable Artificial Intelligence (GeoXAI) system. By addressing limitations in traditional GeoXAI tools, such as the lack of geographical context for model predictions and local explanation, this system integrates state-of-the-art XAI methods with open-source geospatial technologies. Applied to the wildfire-prone regions of Berlin and Brandenburg, Germany, the system combines environmental, topographic, and meteorological features derived from high-resolution geospatial data for training a Random Forest (RF) model. The web-based GeoXAI system enables interactive exploration of the model output and its features, allowing users to visualize wildfire susceptibility, examine feature contributions, and correlate predictions with spatial patterns through post-hoc interpretability. By employing post-hoc explanation methods like SHAP, the system offers clear insights into model predictions by analyzing feature contributions after training, which helps users better understand AI-driven outcomes. Designed with a user-centered approach, the platform promotes trust and usability through transparent predictions, interactive geovisualizations, and local explanations, allowing users to navigate spatial data intuitively by exploring overviews, focusing on specific regions, and accessing detailed insights on demand. This work highlights the potential of combining GeoXAI with machine learning to improve decision-making in wildfire prevention and management.
Life cycle assessments (LCAs) in the construction sector often analyse buildings or their individual components. Applying LCA to determine the environmental impact of entire settlements is less established and the structural infrastructure is often not taken into account. The research project ‘Q-LCA - Analysis of the ecological impacts of different settlement types in new urban development projects over their life cycle’ follows the objective of determining and comparing the material flows, energy consumption including grey energy and the associated emissions of urban settlement components with a focus on global warming potential (GWP). Based on a modular approach, the GWP for 972 different settlement scenarios was determined through an LCA-based model in which six key parameters - building type, settlement density, road layout, building construction method, energy efficiency level and heating system - were systematically combined. The study demonstrates how modularity enables the assessment of large-scale systems such as urban settlements, contributing to an enhanced reflection of their inherent heterogeneity. Besides quantifying the infrastructure share of settlements’ GWP, further results indicate a high influence of operational energy consumption as well as choice of construction method on the area-based GWP. Underground garages are moreover responsible for high GWPs in densified settlements with large building types. This research fills a critical gap in existing literature by emphasising the potential of modular LCA approaches to assess urban development and additionally provides valuable insights for urban planners as well as policy makers seeking to lower environmental impact, identify circularity potentials and mitigate climate change effects in settlements.
Critical Interactivity
(2025)
We propose critical interactivity as a concept to study and design the dynamic and transitory aspects of data visualizations. Theoretically, interactivity is often described as the means to support analytical tasks, while in practice, it encompasses the techniques that alter visual representations. These notions are a useful starting point to study the role of interactivity in critical engagements with data visualizations. At the core of critical interactivity is the negotiation of authority and agency: authority as authors provide structure and context, and agency as viewers navigate and interpret the data on their own terms. This raises the critical question: who has the power to control the visualization? Drawing from four case studies in science communication, art history, anthropology, and climate advocacy, we examine how critical interactivity links exploration and narration. We reflect on the effort involved in preparing data and propose design strategies for implementing critical interactivity in data visualization.
KidRewi ist ein Forschungsprojekt, das sich der Entwicklung und Gestaltung von Publikationsinfrastrukturen für die Rechtswissenschaft widmet. Im Fokus steht die Konzeption einer mit Open-Access- und Open-Science-Prinzipien kompatiblen Infrastruktur. Der folgende Beitrag gibt zunächst einen allgemeinen Überblick über das Projekt und beleuchtet dessen theoretischen Hintergrund. Anschließend werden das methodische Vorgehen sowie die beteiligten Projektpartner detailliert vorgestellt.
CCPL Handbuch und Kommentar
(2025)
Der Beitrag gibt einen Einblick in die Entstehungsgeschichte und die Besonderheiten des ersten offen lizenzierten Rechtskommentars für die Creative Commons-Lizenzen in Version 4.0. Als Projekt von OpenRewi e.V. ist er ein Transformationsbeispiel für die offene Rechtswissenschaft, dem hoffentlich weitere folgen werden.
Trotz der zunehmenden gesellschaftlichen und wissenschaftlichen Auseinandersetzung mit sexualisierter Gewalt in Institutionen bleibt der frühpädagogische Bereich empirisch unterbeleuchtet. Dieser Beitrag analysiert den institutionellen Umgang mit sexualisierter Gewalt in Kitas anhand zweier Fallstudien. Er analysiert die Dynamiken der Thematisierung und Anzeige sexualisierter Gewalt durch Eltern gegen Kita-Personal und externe Personen. Dabei zeigt sich, dass kitaspezifische Strukturen, geprägt durch vielschichtige Abhängigkeitsverhältnisse und institutionelle Geschlossenheit, die Abwehr von Thematisierungen sexualisierter Gewalt begünstigen. Die Untersuchung identifiziert zentrale Abwehrmechanismen, die eine institutionelle Auseinandersetzung erschweren, wie den Verweis auf langjährige kollegiale Beziehungen und die Mystifizierung der Vorwürfe. Die Fallstudien verdeutlichen, dass die institutionellen Reaktionen stärker auf den Schutz der angezeigten Fachkräfte ausgerichtet sind als auf den Kinderschutz. Die Ergebnisse werfen Fragen zur institutionellen Verantwortung und zu den Herausforderungen einer effektiven Aufarbeitung sexualisierter Gewalt in Kitas auf und liefern wertvolle Implikationen für das professionelle Handlungsfeld der Frühpädagogik.
Revealing relevant information on demand is an essential requirement for visual data exploration. In this state-of-the-art report, we review and classify techniques that are inspired by the physical metaphor of un/folding to reveal relevant information or, conversely, to reduce irrelevant information in data visualizations. Similar to focus+context approaches, un/foldable visualizations transform the visual data representation, often between different granularities, in an integrated manner while preserving the overall context. This typically involves switching between different visibility states of data elements or adjusting the graphical abstraction linked by gradual display transitions. We analyze a literature corpus of 101 visualization techniques specifically with respect to their use of the un/folding metaphor. In particular, we consider the type of data, the focus scope and the effect scope, the number of un/folding states, the transformation type, and the controllability and interaction directness of un/folding. The collection of un/foldables is available as an online catalog that includes classic focus+context, semantic zooming, and multi-scale visualizations as well as contemporary un/foldable visualizations. From our literature analysis, we further extract families of un/folding techniques, summarize empirical findings to date, and identify promising research directions for un/foldable data visualization.