A scientometric analysis of 269 articles shows XAI research in decision dashboards shifting from algorithm-centered to human- and decision-centered
Synopsis
Using scientometric methods on 269 original research articles retained after PRISMA screening from Scopus and Web of Science, this study finds that the intellectual core of XAI research in decision dashboards is organized around explainable AI, decision making, and deep learning, that themes have shifted over time from classical machine learning such as neural networks and support vector machines toward visual deep learning, visualization techniques, trust, and intelligent decision support systems, and that the field is moving from a technology-centered toward a human- and decision-centered paradigm.
Interpretation
The field's conceptual network is organized around three central concepts—explainable AI, decision making, and deep learning—and density visualization shows the highest concentration of research at their intersection. Prior work had not systematically examined the intellectual structure of XAI research within decision dashboard design; this study characterizes its conceptual core through keyword co-occurrence networks. Based on 269 original research articles retained after PRISMA screening of Scopus and Web of Science records, analyzed with VOSviewer co-occurrence, density, and overlay visualizations.
The study identifies interconnected thematic clusters including human ethical considerations, technical methodological development, applied operational contexts, cognitive autonomous systems, and emerging specialized topics, reflecting the convergence of algorithmic performance and engineering efficiency with human understanding, trust, and accountability. It integrates what were previously separate technical and human-centered strands into complementary paradigms within a single knowledge structure. Derived from co-occurrence network clustering and thematic evolution mapping, constituting structural evidence at the bibliometric level.
Overlay visualization shows a thematic transition over time: earlier research focused on classical machine learning techniques such as neural networks and support vector machines, while recent studies increasingly emphasize visual deep learning, visualization techniques, trust, and intelligent decision support systems. It presents the field's topic migration along a temporal dimension, noting that trust and visualization have moved from peripheral topics to integral components of XAI systems. Based on time-sliced analysis using VOSviewer overlay visualization and thematic evolution mapping.
Logistic growth modeling indicates the field is still in a rapid expansion phase and has not yet reached scientific saturation. It offers a quantitative perspective on the field's developmental stage rather than relying on the intuitive impression of publication counts. Conclusion drawn from fitting a logistic growth model to the literature growth trajectory.
Perspective
The study targets researchers and design practitioners concerned with explainable AI, decision support systems, and visualization interface design, and applies to contexts where bibliometric methods are used to understand the knowledge structure of this interdisciplinary field; its conclusions describe the distribution and evolution of research themes rather than the effectiveness of any specific dashboard design.
This is an incomplete reading lacking figures, keyword lists, and references, so the specific composition of each thematic cluster, the concrete results of Bradford's and Lotka's laws, and the parameters of the logistic growth model cannot be verified; moreover, bibliometric results reflect the distribution of published literature, and their stability over time remains to be observed in future research.