Data in context: How digital transformation can support human reasoning in cyber-physical production systems

In traditional production plants, current technologies do not provide sufficient context to support information integration and interpretation. Digital transformation technologies have the potential to support contextualization, but it is unclear how this can be achieved. The present article reviews psychological literature in four areas relevant to contextualization: information sampling, integration, categorization, and causal reasoning. Characteristic biases and limitations of human information processing are discussed. Based on this literature, we derive functional requirements for digital transformation technologies, focusing on the cognitive activities they should support. We then present a selection of technologies that have the potential to foster contextualization. These technologies enable the modelling of system relations, the integration of data from different sources, and the connection of the present situation with historical data. We illustrate how these technologies can support contextual reasoning and highlight challenges that should be addressed when designing human-technology cooperation in cyber-physical production systems.

💡 Research Summary

**
The paper investigates how digital transformation technologies can support human reasoning in cyber‑physical production systems (CPPS), focusing on the problem of contextualizing the massive and fragmented data that operators must handle. It begins by describing the current situation in traditional plants: operators are inundated with alarms (often dozens per hour) and raw sensor values, while crucial contextual information—such as data from preceding process steps or environmental conditions—is frequently unavailable. This creates a “context‑dependence” problem that the authors break down into four cognitive activities derived from psychological research: (1) sampling the available information, (2) integrating heterogeneous information elements, (3) categorizing objects and events, and (4) reasoning about causes.

The authors review empirical findings from four areas of cognitive psychology. In the sampling stage, they highlight that people are good at drawing inferences from a given sample but poor at judging its representativeness, leading to sampling bias, availability heuristics, conditional sampling, and confirmation bias. Task complexity and information overload exacerbate these biases, causing operators to focus on salient cues (e.g., temperature spikes) while ignoring less obvious but relevant variables. In the integration stage, the heterogeneity of data sources (different formats, timestamps, and levels of abstraction) creates additional cognitive load; operators tend to use information that is most accessible, even when it is not the most informative. The categorization stage suffers from a lack of explicit situation models, forcing operators to rely on personal heuristics that can lead to over‑generalization and category bias. Finally, causal reasoning is hampered by the complex interdependencies of physical laws, process parameters, and external factors; humans often simplify causal chains or over‑trust automated suggestions, which can mask hidden failure modes.

From these psychological insights, the paper derives functional requirements for digital transformation technologies. The authors argue that technologies must (a) make the sampling process transparent, (b) provide semantic links between disparate data sources, (c) support dynamic, hierarchical categorization, and (d) enable explicit causal modeling and simulation. They then survey a selection of technologies that can meet these requirements:

1. Ontologies and Linked Data – provide a formal, machine‑readable representation of system relations, allowing automatic mapping of heterogeneous data and facilitating context‑aware queries.
2. OPC UA (Open Platform Communications Unified Architecture) – a standardized industrial communication protocol that integrates real‑time data streams with rich metadata, ensuring secure and interoperable data exchange.
3. Data Lakes with Semantic Layers – store raw data from multiple sources while a semantic overlay enables meaning‑based retrieval and fusion.
4. Digital Twins – create virtual replicas of physical assets that combine historical and live data, supporting situation awareness, “what‑if” analysis, and causal inference.
5. AR/VR‑based Visualization – present complex, multi‑source information in intuitive visual or immersive formats, reducing cognitive load and improving pattern recognition.

The authors discuss challenges that arise when deploying these technologies. First, bias transfer: automated support can reinforce existing human biases (e.g., confirmation bias) if not designed with counter‑bias mechanisms. Second, transparency and explainability are essential; operators must understand why a system recommends a particular alarm or action to maintain trust. Third, security and privacy concerns emerge when large volumes of sensor data are shared across networks. Fourth, human‑machine collaboration design must balance automation with operator control, providing timely, filtered information and clear feedback loops. The paper recommends design principles such as adaptive information filtering, user‑customizable views, and training programs that familiarize operators with the underlying data models and causal graphs.

In conclusion, the study positions digital transformation technologies as enablers rather than sources of new problems for CPPS operators. By grounding technology requirements in well‑established cognitive psychology findings, the authors demonstrate that ontologies, standardized communication, data integration platforms, digital twins, and immersive visualizations can collectively mitigate sampling bias, integration overload, categorization errors, and causal misinterpretation. Nevertheless, successful implementation demands careful attention to bias mitigation, explainability, security, and user‑centered design, fostering a symbiotic relationship where technology extends human reasoning without eclipsing it.