How we know

Every figure on this site comes from a public source and is processed by scripts anyone can rerun. This page explains how each answer is worked out and where its limits are. Credits and licences are on the sources page.

The jobs

The jobs are the occupations of ESCO, the European Commission's classification of occupations and skills (version 1.2.1). Each of its 1,791 main occupations has a page; its 1,248 more specialised occupations (for example blockchain developer, under software developer) are listed on their main job's page and found by the search. ESCO links every occupation to a group of the international classification ISCO-08, and most figures come at group level: when a figure describes a wider group than the job itself, a grey tag names that group.

Will there be jobs?

• Jobs to fill by 2035: from Cedefop's Skills Forecast (2026 edition) for the job family (ISCO-08 sub-major group): new jobs from growth plus the people leaving the work who need replacing, 2026 to 2035, per 100 people in the work in 2025. The forecast's measured history ends in 2023. Very small groups in a country are marked as rough.
• Right now: each country's assessment of shortage and surplus occupations for the EURES network (2025 round), for the narrower unit group (ISCO-08 four digits). These are national expert judgements; their severity can't be compared across countries.
• Advertised: the job group's rank among job ads published on the EURES portal in that country (July 2025 to June 2026). EURES covers each country's job market to a very different extent, so ads are only compared within a country.
• The short answer: more than 1.2 times the all-jobs rate of jobs to fill, or a shortage reported now, gives "Yes, good prospects"; 0.8 to 1.2 times gives "Average prospects"; lower gives "Limited prospects". In the EU view, a shortage means at least 9 of the 27 countries report one.

What does it pay?

• Source: Eurostat's Structure of Earnings Survey 2022, which publishes average pay only for the ten broad occupation groups (ISCO-08 one digit).
• Our estimate for the job family: the broad group's official average is split across its families with the occupation and sector factors of Eurostat's salary calculator, weighted by employment so that the families add back up to the official figure. Pay is then brought up to 2025 with the growth in wages (labour cost index). It is an estimate, labelled as such.
• Under 30 and the gender pay gap: pay of under-30s comes from the same survey; the gender pay gap is Eurostat's own figure (difference in average hourly pay) for the broad group.
• The short answer: more than 40% above the country's all-jobs average is "well above average", 20 to 40% above is "above average", within 20% either way "about average", lower "below average". In the EU view, pay differs too much between countries for one figure: the page shows the range, and the short answer uses the middle value of the countries' comparisons.

What to study

• Usual level: the ILO skill level of the job family: level 4 is "usually a degree", 3 "usually post-secondary or a short degree", 2 "upper secondary or vocational", 1 "basic schooling". The qualifications people in the family hold now come from Cedefop's forecast.
• Fields of study for a job: ESCO files every knowledge item under a field of education (ISCED-F 2013). A job's fields are the fields of the knowledge ESCO lists for it, blended with its unit group's when the job lists little knowledge. A field is shown only when the evidence is solid; otherwise the page shows the study level only. These are fields that teach what the job uses, not a record of what people in the job studied.
• Where a field can lead: the same link the other way round: the jobs that use what the field teaches. "What you'd learn" lists the field's knowledge items most needed by jobs.
• Vocational or degree: Eurostat counts of graduates per field and level in each country (latest year, mostly 2023 or 2024). Where a country reports vocational graduates only for a wider field, that field is used and the page says so.
• Do graduates use it? The share of 15 to 34-year-olds whose job matches their field of education, from the 2024 module of the EU Labour Force Survey, by broad field (for example all ICT fields together). Eurostat marks some of these figures as unreliable because samples are small; the pages say "small sample".
• Do they find work? Eurostat's employment rate of recent graduates: people aged 20 to 34 who finished education 1 to 3 years earlier, by level and route, for all fields together.
• Where to study: the European Tertiary Education Register (ETER): higher-education institutions still listed in their country's latest data, ranked by students in the broad field (ETER has no finer field detail). For vocational routes and for the actual programmes, the site links each country's official course finders, a list we compiled and check.

Will AI change it?

The International Labour Organization's 2025 index of exposure to generative AI scores every task of each unit group from 0 (current AI can't do it) to 1 (it could do all of it), then places the group in one of six categories, which the site names very low, low, some tasks, moderate, significant and high. It describes what AI could technically do when assessed in 2025: it is not a forecast for a future year, and not a forecast of job losses. Unit groups the ILO does not score show the average of related groups, labelled as such.

What the work is like (US data)

No EU source describes working conditions per job in this detail, so this part uses O*NET 31.0, the US Department of Labor's database, and always names the US job it comes from. Each ESCO job is matched to one US job by a fixed rule (same title first, then the European Commission's crosswalk), and seven averages from US workers (where, sitting or standing, contact with people, freedom to decide, usual weekly hours, deadlines, repeating the same tasks) are turned into short phrases by fixed thresholds. Repetition is only described at the clear ends, because the middle of that scale is unreliable.

Search and areas of work

The search looks through every ESCO job name, alternative name and specialisation in English, and corrects likely typos against the words ESCO uses. The 18 areas of work on the start page are our own grouping of ISCO-08 groups, in plain words; the table is published with the data and open to review.

Updates and versions

Each source is stored exactly as downloaded, with its date and a fingerprint, and every figure on the site is rebuilt from those files. Main versions: ESCO 1.2.1, Cedefop Skills Forecast 2026, EURES shortages 2025 round, Structure of Earnings Survey 2022 (pay brought to 2025), ILO generative AI index 2025, O*NET 31.0, ETER 2011 to 2024, Eurostat education and labour force tables downloaded in November 2026.