Zhaolu Liu,Jonathan M. Clarke,Bertha Rohenkohl,Mauricio Barahona
Abstract:A job usually involves the application of several complementary or synergistic skills to perform its required tasks. Such relationships are implicitly recognised by employers in the skills they demand when recruiting new employees. Here we construct a skills network based on their co-occurrence in a national level data set of 65 million job postings from the UK spanning 2016 to 2022. We then apply multiscale graph-based community detection to obtain data-driven skill clusters at different levels of resolution that reveal a modular structure across scales. Skill clusters display diverse levels of demand and occupy varying roles within the skills network: some have broad reach across the network (high closeness centrality) while others have higher levels of within-cluster containment, yet with high interconnection across clusters and no skill silos. The skill clusters also display varying levels of semantic similarity, highlighting the difference between co-occurrence in adverts and intrinsic thematic consistency. Clear geographic variation is evident in the demand for each skill cluster across the UK, broadly reflecting the industrial characteristics of each region, e.g., London appears as an outlier as an international hub for finance, education and business. Comparison of data from 2016 and 2022 reveals employers are demanding a broader range of skills over time, with more adverts featuring skills spanning different clusters. We also show that our data-driven clusters differ from expert-authored categorisations of skills, indicating that important relationships between skills are not captured by expert assessment alone.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to understand the relationships between different skills in the UK labor market and their changes over time. Specifically, by analyzing the skill co - occurrence patterns in 65 million UK job advertisements from 2016 to 2022, the author constructs a skill network and uses the multi - scale graph clustering method to reveal skill clusters at different resolutions. The main research questions include:
1. **Identification and characteristics of skill clusters**: Identify skill clusters with similar co - occurrence patterns based on the co - occurrence frequency of skills in the network, and analyze the characteristics of these clusters, such as centrality, semantic similarity, and self - containment of skills within the cluster.
2. **Spatial - temporal changes in skill demand**: Explore the differences in the demand for different skill clusters across the UK and their changing trends over time. For example, London's position as an international financial, educational, and commercial hub, and the unique skill demand patterns in other regions.
3. **Comparison between skill clusters and expert classifications**: Compare the data - driven skill clusters with the skill classifications written by experts to reveal important relationships between skills that expert evaluations have failed to capture.
4. **Skill complementarity and synergy effects**: Study the comprehensive demand of employers for multiple skills and the interactions of these skills in the network, especially the relationships between cross - cluster skills.
### Main research content
- **Data sources and processing**: Use 65 million UK job advertisement data collected by Adzuna Intelligence, covering 2016 - 2022. Each advertisement contains the release date, geographical location, and at least one skill.
- **Construction of skill network**: Construct a weighted undirected similarity graph \(G\) through the co - occurrence of skills in advertisements, where nodes represent skills and the edge weights represent the similarity of skill co - occurrence.
- **Multi - scale clustering analysis**: Apply the Markov Stability (MS) method for multi - scale graph clustering to extract skill clusters at different resolutions. Focus on analyzing 21 skill clusters at medium resolution (MS21) and describe their characteristics in detail.
- **Analysis of network properties**: Calculate properties such as centrality, semantic similarity, and self - containment of skill clusters to evaluate the role and importance of skills in the network.
- **Analysis in geographical and temporal dimensions**: Analyze the distribution of different skill clusters in various regions of the UK and their changing trends over time, revealing regional economic characteristics and the dynamic changes in skill demand.
### Conclusion
The study found that the skill demand in the UK labor market is highly diverse and interconnected, and different skill clusters play different roles in the network. Some skill clusters have extensive network connections, while others are more self - contained but have strong cross - cluster connections with other clusters. In addition, skill demand shows significant differences in different regions and time periods, reflecting the industrial characteristics and economic structures of each region. Overall, over time, employers' demand for diverse skill combinations has gradually increased, indicating that the modern labor market has an increasing requirement for comprehensive skills.