1 Introduction

1.1 Motivation

Returning to academia after a significant period in the industry provides a fresh perspective to examine how knowledge management intersects with education. The selected timeframe (2023–2024) aligns with the accelerating trend towards data-driven decision-making in education, where the need for rapid knowledge transformation and real-time application has become critical, as noted by [17] in the context of data-centric educational frameworks. This shift opens up an exciting chance to closely examine how the SECI model is applied in Western educational settings, an area that hasn't been thoroughly explored despite its success in Eastern business environments and its potential global relevance. As Farnese et al. [1] point out, there's a notable lack of detailed study in this area, highlighting an important gap in our understanding of the model's flexibility and practical use. The investigation was concluded by developing and implementing one course and one three-course-micro-degree program based on the SECI model to facilitate knowledge transformation. The study included a diverse sample of 32 learners from diverse demographic backgrounds (geography: Estonia, Finland, Turkey, France, Cameroon, Nigeria, India; education-level: from Bachelor to Master Degree; industry: finance, manufacturing, game design, software development, Fast-moving consumer goods, e-commerce, public sector), what enhanced the representativeness of the findings. Leveraging a strong background in Computer Science, the approach will use data analysis and pattern recognition, to evaluate the SECI model's effectiveness, while employing measures such as triangulation (data sources from multiple feedback surveys from multiple courses; three different course content files, field observations), to mitigate potential biases in the feedback data. This way, the study combines educational research with data science techniques.

1.2 SECI model for education

In the evolving landscape of digital education, the SECI model, as elucidated by [11], presents a structured framework for the transformation of tacit knowledge into explicit knowledge, thereby facilitating a dynamic learning environment. The model's integration into course design, particularly through the improvement of problem-based learning (PBL) approaches, enhances knowledge transfer, engagement, and practical application among learners, as demonstrated by [15]. Furthermore, the importance of aligning educational practices with industry demands, highlighted by [7], underscores the model's applicability in preparing students for professional challenges. In [17] the updated model was introduced, which takes into consideration specifically student’s and lecturer’s experiences, and feedback loop, placed in the realm of the SECI model.

This study, drawing upon insights from [2] and [3], explores the versatility of the SECI model across various educational fields, emphasizing its potential to bridge educational practices with pragmatic professional world demands. By evaluating the SECI model's application in recent course developments, this research aims to uncover effective strategies that resonate with learners and promote deep learning. Despite the well-established theoretical benefits of the SECI model in knowledge management, its empirical examination within the context of digital education, particularly in creating engaging learning environments, remains limited. The facilitation of tacit knowledge sharing is augmented by insights from [15] and [4], who discuss the complexities of knowledge codification and transfer in higher education. This gap is addressed by evaluating course designs and feedback to discern patterns indicative of successful knowledge transformation, thus validating the hypothesis that courses rooted in the SECI model's principles will garner positive feedback from participants and demonstrate effective knowledge transfer.

2 Methodology

In this study, the author utilizes a combination of qualitative and quantitative methodologies to examine the application and results of the SECI model in the developed courses. Central to the investigation is the use of natural language processing (NLP) and text analysis to analyze course content and learner feedback from the ‘Innovation and Digitization Management’ and ‘Data-Based Decision Making’ programs to uncover recurring themes and patterns that correlate with positive learning outcomes.

The research involves a detailed collection of data from a diverse sample of 32 students, including course materials, educational frameworks, and anonymous feedback collected in 7 surveys, providing a solid foundation for subsequent analysis.

The approach adopted is data-centric, consisting of four steps:

  1. 1.

    Data preprocessing. The initial step involves an exhaustive cleaning and organization of text data ensuring the removal of any irrelevant or biased information..

  2. 2.

    Tokenization. Breaking down text into individual terms or phrases.

  3. 3.

    Vectorization. Converting text data into numerical format using TF-IDF (Term Frequency-Inverse Document Frequency) to reflect the importance of words within the documents.

  4. 4.

    Clustering. Implementing K-means clustering to group similar text data, facilitates the identification of dominant themes and concepts in the course content, thereby offering insights into how the course content aligns with the SECI model's principles.

Challenges encountered during data analysis, such as dealing with noisy data and ensuring the accuracy of clustering results, were addressed by using robust data preprocessing techniques and validating clustering outcomes through multiple iterations.

This methodological approach, supported by [8] and [13], allows for an in-depth understanding of how the SECI model is applied in digital education courses and its impact on learner engagement and knowledge acquisition. Additionally, the integration of socialization, externalization, and feedback loops as discussed in [17] provides a qualitative depth to the current study, balancing the quantitative findings with critical insights into the tacit-to-explicit knowledge transformation processes.

The findings aim to provide actionable insights for educators and instructional designers seeking to integrate knowledge management frameworks into digital learning environments. Inspired by the efficiency of cross-lingual alignment techniques [9], our approach seeks to extend the analytical depth by considering the nuances of language and terminology used across different educational contexts.

3 Findings

3.1 Course feedback from participants

Feedback underscores the practical application of the SECI model in enhancing learning experiences, reflecting the Socialization and Externalization phases, enriched by referencing studies such as [15]. This alignment with the SECI model phases is enriched by empirical evidence, illustrating a comprehensive view of the feedback's alignment with the SECI model phases.

Suggestions for clearer documentation and more examples align with the combination phase, aiming for better structured and systematic knowledge. The call for real-world applications and guest speakers underscores the internalization phase, where explicit knowledge is transformed into tacit understanding, highlighting the SECI model's comprehensive impact on educational outcomes. Together, these feedback surveys present a comprehensive view of the feedback across the three courses (data-based decision-making process, data-based decision-making leadership, innovation and digitization management), underscoring the value of practical, interactive learning and the effective use of the SECI model, alongside areas for enhancing course delivery and content clarity.

To further deepen our understanding of learner experiences and perceptions, future research could incorporate additional qualitative methods such as in-depth interviews and focus groups. These methods would provide richer, more nuanced insights into learner experiences, complementing the quantitative analysis and enhancing the overall robustness of the findings.

3.2 Text analysis

The K-means clustering analysis yielded four clusters, each representing a topic concentration. This analysis highlights the curriculum's diversity, including practical tools, techniques, strategies, and case studies [8]. Cluster 0 focuses on pedagogical strategies, echoing the importance of innovative teaching and information sources [6, 14, 15]. Clusters 1 and 2 emphasize analytical methods and structured educational research. Cluster 3, highlighting real-world applications, aligns with explorations of the SECI and Leader-member exchange (LMX) theory in workplace preparedness [5].

These results illustrate the diverse thematic focuses across the courses, showcasing a comprehensive approach to incorporating data analytics, technology, and pedagogical strategies in the curriculum, aligning with the study's findings on the beneficial impact of such integration on teaching practices and educational outcomes (Table 1).

Table 1 Distinct clusters representing a concentration of topics
Full size table

This study further explored the thematic structure of course materials through advanced text analysis, employing Latent Dirichlet allocation (LDA) for topic modeling and creating a diagram to map identified themes to the SECI model phases. After the clustering, we embarked on a comparative document analysis to evaluate the similarities between the course contents, employing three different mathematical approaches: cosine similarity, Euclidean distance similarity, and Jaccard similarity.

Cosine similarity calculations revealed notable associations between documents, particularly between those that were thematically coherent. Documents within the same cluster showed higher degrees of similarity, which was expected given their topical alignment. Specifically, the documents grouped in Clusters 0 and 3 demonstrated a higher cosine similarity score, suggesting a closer thematic relationship, perhaps due to shared jargon or overlapping subject matter.

Euclidean distance similarity, when inverted to form a similarity measure, provides a nuanced understanding of document relatedness. This metric highlights the differential spacing between documents in a multi-dimensional space, offering a perspective that considers the magnitude of term frequencies. In this analysis, Clusters 1 and 2 showcased the largest distances, alluding to distinctive content that sets them apart from other clusters.

Jaccard similarity, which is sensitive to the size of the document, as it measures the proportion of shared terms, provides a more stringent measure of similarity. This binary-based measure underscores the shared vocabulary across documents, revealing an intriguing interplay of commonality and uniqueness within the course material. The documents within Clusters 0 and 3 shared a greater proportion of terms, reinforcing the insights gained from the cosine similarity analysis.

These computational techniques paint a comprehensive picture of the textual landscape of the course materials. While Clusters 0 and 3 shared a significant overlap in terms, indicative of related pedagogical strategies or conceptual frameworks, Clusters 1 and 2 were characterized by their distinctiveness, which could be attributed to specialized content unique to the particular courses they represented.

Heatmaps of these similarity measures were visualized, providing a vivid illustration of the inter-document relationships (Fig. 1). The heatmaps served as a testament to the thematic richness and diversity within the courses and flagged potential areas for content integration and inter-course connectivity. The visual analytics further supported the SECI model's phases, underscoring the socialization and externalization in the shared knowledge of Clusters 0 and 3 and the combination and internalization in the more distinct knowledge areas of Clusters 1 and 2.

Fig. 1
figure 1

Heatmaps for clusters—cosine similarity, Euclidean distance, and Jaccard similarity

Full size image

This dual approach aims to elucidate the comprehensive integration of socialization, externalization, combination, and internalization processes within the curriculum. By visually representing these alignments, we can pinpoint both the strengths of facilitating a holistic learning experience and the areas ripe for enhancement to deepen the application of the SECI model in educational settings.

4 Conclusions

This study contributes to the field by demonstrating the practical application of the SECI model in digital education and offering a nuanced understanding of the interaction between tacit and explicit knowledge through a mixed-methods approach. This dual focus ensures that the findings are robust and applicable across diverse educational contexts. In line with our ongoing efforts to enhance digital education frameworks, this study builds upon our previous work [17], exploring the integration of the SECI model and digital innovation to advance knowledge transformation in MBA education. Our findings underscore the significance of adopting innovative teaching methodologies and leveraging digital tools to facilitate effective knowledge transfer, thereby enhancing the learning experience and preparing students for professional challenges. The clustering results reveal a balanced integration of theory and application within the course materials, aligning well with the SECI model's phases of socialization, externalization, combination, and internalization.

The presence of diverse themes—from pedagogical methods to applied analytics—suggests a comprehensive approach to knowledge transformation, encouraging active engagement and practical application. Our analysis showcases a robust alignment of course content with the SECI model's phases, promoting a dynamic learning environment that encourages practical application. The cultural transmission of tacit knowledge [10] and the efficacy of flipped classrooms [12] highlight the evolving pedagogical strategies essential for modern education. Furthermore, the adaptability and relevance of the SECI model in current educational contexts, as discussed by [16], underscores its potential for continuous evolution to meet the challenges of contemporary education.