occupation_key stringlengths 8 21 | occupation_label stringlengths 8 32 | ai_exposure_class stringclasses 2
values | world_count_change stringclasses 2
values | direction_status stringclasses 3
values | markets_visible int64 7 24 | markets_with_fewer_postings int64 0 21 |
|---|---|---|---|---|---|---|
accounting_finance | Accounting & bookkeeping | AI-exposed | rose | published (robust) | 22 | 7 |
admin_clerical | Administrative & clerical | AI-exposed | rose | not published | 20 | 3 |
agriculture | Agriculture & landscaping | AI-resilient | rose | published (robust) | 8 | 1 |
sales_b2b | B2B / field sales | AI-exposed | rose | published (robust) | 21 | 10 |
personal_services | Cleaning, food & security | AI-resilient | rose | published (robust) | 17 | 0 |
construction | Construction & building trades | AI-resilient | rose | published (robust) | 12 | 1 |
customer_support | Customer support & call center | AI-exposed | rose | published (robust) | 19 | 6 |
data_analytics | Data analysts & entry | AI-exposed | fell | published (robust) | 21 | 8 |
transport_driver | Drivers & transport | AI-resilient | rose | published (robust) | 16 | 1 |
electrical | Electricians | AI-resilient | rose | not published | 14 | 3 |
legal_support | Legal support | AI-exposed | rose | not published | 8 | 2 |
marketing_content | Marketing, content & translation | AI-exposed | fell | published (robust) | 18 | 13 |
mechanic_maintenance | Mechanics & maintenance techs | AI-resilient | rose | published (sensitive) | 18 | 4 |
healthcare_frontline | Nurses & care workers | AI-resilient | rose | published (robust) | 14 | 1 |
healthcare_physicians | Physicians & doctors | AI-resilient | rose | published (robust) | 11 | 1 |
plumbing | Plumbers & pipefitters | AI-resilient | rose | published (robust) | 7 | 1 |
production_operator | Production & machine operators | AI-resilient | rose | published (robust) | 12 | 2 |
real_estate_broker | Real-estate agents/brokers | AI-resilient | rose | not published | 9 | 4 |
retail_sales | Retail & shop assistants | AI-resilient | rose | published (robust) | 22 | 1 |
software_dev | Software developers | AI-exposed | fell | published (robust) | 24 | 21 |
teaching | Teachers | AI-resilient | rose | published (robust) | 17 | 2 |
warehouse_logistics | Warehouse & logistics | AI-resilient | rose | published (robust) | 11 | 1 |
welding_metal | Welders & metalworkers | AI-resilient | rose | published (sensitive) | 13 | 3 |
SHIFT Labour Demand Index 2024–2026 (Edition 1)
The occupation margin of the SHIFT Labour Demand Index. What happened to advertised demand for 23 kinds of work across 24 markets, between the twelve months to June 2024 (4,241,295 advertisements counted) and the twelve months to June 2026 (9,323,023 advertisements counted).
It is a count of what employers already advertised for — not a forecast, not a survey, and not a sentiment index. This public file reports the worldwide direction of the raw advertisement count for each kind of work — whether it rose or fell in absolute terms — not a change in share. The index's market-relative measure (the SHIFT Pull — how an occupation moved against its own national market) is the paid product and is not included in this file.
This public file is the occupation margin: the direction of change and how widely it held across markets. It does not include the underlying magnitudes or the per-market detail, which belong to the fuller SHIFT dataset.
What's in this dataset
A single table, shift_ldi_2026_e1_by_occupation.csv — one row per occupation (23 rows).
| Column | Meaning |
|---|---|
occupation_key |
Stable machine key for the occupation. |
occupation_label |
Human-readable occupation name. |
ai_exposure_class |
Whether the occupation is broadly AI-exposed or AI-resilient. |
world_count_change |
Direction of change in advertised demand across visible markets: rose or fell. |
direction_status |
Confidence/publication status of that direction: published (robust), published (sensitive), or not published. |
markets_visible |
How many of the 24 markets had enough data for the occupation to be visible. |
markets_with_fewer_postings |
Of those visible markets, how many saw advertised demand fall. |
Coverage
- 23 occupations spanning AI-exposed and AI-resilient work (software, data, accounting, marketing, customer support, legal and admin alongside trades, healthcare, logistics, construction, retail and teaching).
- 24 markets, 9 languages.
- Two windows: the 12 months to June 2024 vs the 12 months to June 2026.
A quick read of Edition 1
Advertised demand rose for most kinds of work over the period, but it fell for several highly AI-exposed desk occupations — software developers, data analysts, and marketing / content / translation — and those declines were the most widespread across markets (software fell in 21 of 24 visible markets). The trades, healthcare and logistics were broadly resilient. The dataset lets you check this per occupation rather than taking a headline on trust.
How to cite
Covas, J. (2026). SHIFT Labour Demand Index 2024–2026 (Edition 1) [Data set]. SHIFT Research. https://doi.org/10.5281/zenodo.22346056
Links & provenance
- Home: https://shiftindex.org/ · How to cite: https://shiftindex.org/cite
- DOI (this version): https://doi.org/10.5281/zenodo.22346056 · Concept DOI: https://doi.org/10.5281/zenodo.22346055
- Source repository: https://github.com/Shiftindex-org/shift-labour-demand-index
- Query it live (API / agent tool): https://api.shiftindex.org
- Full methodology & data dictionary: see the Zenodo record above.
Creator
José Covas — founder of SHIFT Research. Wikidata: Q141324073 · ORCID: 0009-0000-2298-0626
License
Released under Creative Commons Attribution 4.0 International (CC BY 4.0). You may share and adapt the data, including commercially, with attribution as above.
- Downloads last month
- 47