Nearly half of engineering jobs in the wind industry already request advanced digital skills, with scientific programming and numerical modelling the clearest priorities, while demand for machine learning, IoT and cybersecurity depends on firm and role — signalling a widening skills gap that education providers must address.

Main Finding

A triangulated, mixed-methods analysis of the wind energy sector finds that advanced digital skills are already core to many engineering roles and are expected to broaden rapidly. Scientific programming and numerical modelling are the clearest cornerstone competencies across surveys, job-postings, and expert interviews. Machine learning, IoT, cloud/data-engineering and cybersecurity are important but their prominence depends on organisational context and role; roughly 44% of engineering-related job adverts in the analysed corpus list advanced digital-skill requirements. The study provides a repeatable taxonomy and NLP-enabled workflow to track evolving skill demand and supports targeted upskilling/reskilling and curriculum redesign.

Key Points

• Triangulation: The study combines three complementary sources — job-posting analysis, an industry survey (n = 108), and semi-structured expert interviews (n = 14) — to reduce single-source bias and capture both current demand and foresight.
• Core skills: Scientific programming and numerical analysis/simulation emerge as fundamental across all data sources.
• Variable prominence: Machine learning & data science, IoT & XR, cloud/data engineering, and cybersecurity show heterogeneous importance depending on role, seniority and company context.
• Penetration in job market: ~44% of engineering-related positions in the analysed job-posting set require advanced digital skills.
• Training needs: Interviews highlight persistent skill gaps and the need for lifelong learning; survey respondents indicate preferred training formats and priorities (ranked weighting used to derive importance).
• Repeatability: The study defines a 12-category advanced-digital-skills taxonomy (derived from the Digital Europe Programme) and applies NLP pipelines to enable periodic re-analysis and monitoring.
• Methodological caveats: Keyword-based NLP, limited survey/interview sample sizes, and potential selection bias (EU/consortium networks, English-language data) mean results are descriptive and exploratory rather than causal.

Data & Methods

• Skill taxonomy: 12 advanced digital skill categories mapped from Digital Europe Programme inputs and tuned for wind energy. Examples include: scientific programming; numerical analysis & simulation; machine learning & data science; generative AI & LLMs; data engineering; cloud computing; IoT & extended reality; robotics & autonomous systems; cybersecurity; HPC; digital research communication; blockchain.
• Occupational mapping: ESCO occupation taxonomy used to classify roles and enable cross-walking between job-postings and education/occupational categories.
• Survey:
  • n = 108 valid responses collected via LinkedIn, conferences and project networks.
  • Respondents ranked up to three priority skill areas (weights: rank1 = 3, rank2 = 2, rank3 = 1) and rated training formats on a 5-point Likert scale.
• Interviews:
  • 14 semi-structured interviews across 5 countries and 7 wind-energy subsectors.
  • Protocol: ~1-hour online sessions, transcription, pre-interview survey; thematic questions on current gaps, training needs, and 5–15-year foresight.
  • Qualitative processing: text pre-processing with spaCy (en_core_web_md); passage selection threshold ≥60 characters and ≥12 tokens to exclude non-informative fragments.
  • Thematic coding: primarily deductive keyword-based approach informed by job-posting keywords; 18 subcategories across two high-level groups (Workforce Development & Advanced Digital Skills).
  • Sentiment analysis: applied SiEBERT (RoBERTa-based) for positive/neutral/negative labeling with confidence scores.
  • Social network analysis (SNA): used to measure topic–respondent connectivity and topic centrality.
• Job-posting analysis:
  • NLP-driven extraction of skill mentions using the same keyword taxonomy; used to compute relative importance and the ~44% figure for engineering roles requiring advanced skills.
  • (Paper does not report full job-posting count in the provided excerpt; methodology emphasises repeatable scraping + keyword mapping.)
• Integration:
  • Relative importance scores computed separately per source (count of mentions / total mentions) to make cross-source comparisons straightforward.
• Limitations noted by authors: keyword matching limitations, need for task-specific annotated datasets, short-term hiring dynamics affecting job ads, and representativeness constraints (sample sourcing from consortium networks and English-language materials).

Implications for AI Economics

• Changing labour demand composition:
  • The wind sector exemplifies the broader digital/AI transition: demand shifts toward hybrid roles combining domain expertise and software/data skills. This supports models where AI complements high-skilled labour, increasing demand and potential wage premiums for workers with combined domain + digital capabilities.
• Skill-biased technical change:
  • Findings are consistent with skill-biased technological change: advanced programming and numerical modelling act as general-purpose complementarities for deploying AI/ML and data systems in engineering contexts. Policymakers and firms should expect rising returns to these skills and growing heterogeneity across occupations.
• Training markets and credentialing:
  • The mismatch between university curricula update cycles and rapid industry demand strengthens the case for modular upskilling (micro-credentials, bootcamps, industry–university short courses). Public funding or incentives for lifelong learning could correct market failures (under-provision of rapid re-skilling).
• Labour supply and mobility:
  • A sizeable share of engineering roles already list advanced-skill requirements, implying competition for talent and potential bottlenecks. This could raise recruitment costs and push firms to invest in internal training or rely on outsourcing.
• Heterogeneity by firm role/size:
  • Variation in prominence of ML, IoT, and cybersecurity suggests heterogeneous investment returns across firms. Economic analyses should model firm-level complementarities (e.g., data availability, asset scale) that determine whether AI/ML investments are profitable and which skills are most valuable.
• Policy and educational implications:
  • Results support targeted curriculum updates (more scientific programming, numerical simulation, data engineering) in engineering and technical degrees and expansion of short-cycle, industry-aligned training. ESCO-aligned taxonomies and repeatable NLP monitoring can help policymakers track labour-market signals in near real-time.
• Research agenda for AI economics:
  • Quantify wage premia and employment growth for hybrid roles in renewables; estimate causal effects of upskilling programmes on placement and productivity; model how diffusion of AI across firms affects sectoral productivity and capital–labour substitution in green industries.
• Measurement & monitoring:
  • The paper demonstrates a practical, repeatable methodology for monitoring skill demand using NLP and ESCO mapping. Economists and labour-market policymakers can adopt such pipelines to produce higher-frequency indicators of AI-related skill demand and to assess the impact of interventions.

If you want, I can: - Extract the 12-category taxonomy as a one-page reference for curriculum design. - Draft suggested curriculum modules (learning outcomes and topics) mapped to the top 4–6 skills identified. - Propose an empirical plan to estimate wage premia for advanced-digital skills in the wind sector (data needs, identification strategy).