ABSTRACT To address the complexity of career decision-making limitations, we propose a novel framework for calculating occupation similarity measures based on bipartite graphs constructed from public occupation-skills ontologies. The proposed occupation similarity measures were constructed exclusively utilising knowledge from the European standard classification of occupations (ESCO) ontology. The resulting occupation similarity measures are fully explainable and computationally efficient. Furthermore, they are are effortlessly transferable across regions and countries. This approach allows us to tailor the similarity measures to the specific needs, preferences, and career status of job seekers, providing a more personalised and comprehensive view of potential career paths. Our validation using an extensive dataset of over 450,000 job transitions in Slovenia confirms the effectiveness of our approach, demonstrating the value of employing multiple occupation similarity measures over a single measure.