Only 20% of the world's data is public.
We work entirely with that — no personal or company data.
Above water20%
Below water80%
Job ads
Research papers
Investments
Open data
News
Patents
Grants
Conferences
Personal data
Salary records
Performance reviews
Proprietary data
Internal HR data
Health records
NO PERSONAL DATA
Different formats, contexts, vocabularies — one shared structure.
Job ads recruiter signals
"5+ yrs Python" · "LLM fine-tuning" · "stakeholder mgmt"
Research papers academic terms
"transformer attention" · "few-shot" · "ablation study"
Investments VC taxonomy
"Series B · SaaS" · "vertical AI" · "agentic workflows"
ONTOLOGYShared
structure
structure
UNIFIED KNOWLEDGE GRAPH · one structure · one vocabulary
What each source tells us
Job ads
→
what employers are actually asking for
Research papers
→
where the science is heading
Investments
→
where the money & confidence are going
Open news
→
the wider signal around them
Brought into the same semantic layer — read side by side.
Detailed words — the small items
The data has a catch
Cleaning the job ads
Generative AI floods job ads with "data" and "digital." We keep only the ones genuinely about advanced digital skills.
Filter ·
advanced digital skills
advanced digital skills
✓ kept
✕ filtered out
Advanced digital skills
generic "data"
Machine-learning engineer
AI-generated filler
Cloud security
generic "digital"
Data engineering
keyword-stuffed ad
Cybersecurity analyst
vague "tech"
Only genuine advanced-digital job ads feed the numbers — that's why 2025–26 is higher, but realistic.
What the data tells us
0job ads
0research papers
0investments
→
0jobs · advanced digital skills
Scaled to Europe · Eurostat & Cedefop vacancy rates
≈ 500,000
roles needed in advanced digital skills
…and that becomes the dashboard
