DS Experimental statistics > Online job advertisement rate > Block 1 EN REVAMP

Why are these statistics needed?

Since several years, the European economy has been marked by difficulties for employers to fill their vacant posts. These new experimental statistics based on web sources (online job advertisements) aim to help policy makers, employers, and jobseekers knowing which sectors of the economy might need more labour force.

The existing job vacancy statistics inform users which EU countries and industries, based on the statistical classification of economic activities (NACE), have ‘tight’ labour markets, meaning that there is a relatively high number of vacant posts in comparison with occupied posts.

However, official job vacancy statistics fall short of providing necessary breakdowns by occupation for a detailed assessment of the bottlenecks employers face when recruiting staff. Covering this gap through surveys is not envisaged due to cost and burden.

Why are they experimental?

These statistics are experimental because they are partly compiled from a new data source, the web scraping of online job advertisements ( OJAs) posted on the internet, using natural language processing.

OJAs provide detailed information on:

  • the positions employers are seeking to fill
  • the skills sought for those jobs
  • other characteristics of the job offered and the employer, such as location, sector of economic activity, type of contract, etc.

The dataset has a large coverage in terms of job postings as web-scraping collects all OJAs referenced by web portals selected by the web intelligence hub (WIH).

However, the coverage of the web portals is not exhaustive, meaning the web scraping does not cover all potential web portals posting job advertisements. Thus, the selection of web portals might have a significant impact on the quality of the data output and the comparability across time and countries.

Moreover, the OJA source is biased regarding economic activities and occupations: certain positions are likely to be advertised online more than others. For example, IT positions are more frequently posted on the internet than jobs in small shops, such as butchers or bakeries, or in the public sector (e.g. teachers, care workers). This may lead to the under-coverage of some occupations.

Finally, the same advertisement may be counted multiple times in case it was posted in several web portals. However, the algorithms are progressively improved to eliminate such redundant OJAs.

How are they produced?

The online job advertisement rate (OJAR) is calculated as the ratio of the number of OJAs (from the web intelligence hub - WIH) divided by the number of employees (from the EU-LFS).

The number of OJAs is measured as the total flow of new OJAs posted over a reference period. More information can be found in the metadata on  OJAs.

The number of employees is obtained from the annual  labour force survey (EU-LFS). This survey does not cover persons doing military or community service nor persons living in institutional or collective households.

The online job advertisement rate is computed for each occupation at the 3-digit level of the ISCO classification.

For further details, please consult our  methodological note.

Access to the data

Data navigation tree

DS Experimental statistics > Online job advertisement rate > Block 2 EN REVAMP

Feedback    

To help Eurostat improve these experimental statistics, users and researchers are kindly invited to give us their feedback by email:

  • How useful do you consider statistics on online job advertisement rates?
  • What are your main uses of the statistics on online job advertisement rates?
  • Would you have any comments on the methodology?
  • Would you have any comments on the breakdowns proposed (EU totals, by ISCO-3d occupation)?