Abstract
Digitalisation is assumed to have far reaching consequences for workers. So far, these have been analysed using indicators derived from survey data on occupational tasks. Survey-based indicators measure what people do at work but provide little insight into the skills required to perform a task. Since multiple skills may be necessary to perform a task, approximating digital skills through tasks may underestimate the extent of digitalisation of a given occupation. Besides, they provide limited coverage in terms of periodicity, scope and variety of tasks. We therefore suggest to change the perspective from tasks to skills and propose to analyse the digital skill requirements of occupations. To this end, we use detailed information on the classification of European Occupations, Skills and Qualifications, natural language processing tools and network analysis methods to determine digital skills in the database. We construct four different versions of the digital competencies indicator identifying occupations that depend highly on digital skills. Our indicator can be mapped to the ISCO-08 classification and easily be used alongside other data sources. We show that compared to an indicator based on ICT-tasks derived from the OECD 'Programme for the Assessment of Adult Skills', our indicator captures more complex and specialised digitalised occupations. Our results stress the importance of using granular data in order to properly identify digital skill requirements of jobs.