Alex Lockey Let’s talk

Data analysts

Associate professional occupations, SOC 2020 code 3544. Back to the map

Data analysts gather and organise a variety of data and analyse it to understand what it means for their organisation or society.

Employees (UK)
68k
Median annual pay
£38,107
Exposure score ?
6.4/10 High 9.3/10 Very high strict reading, with tools is 9.3/10 with-tools reading, strict is 6.4/10
Wage exposure
£1.66bn £2.41bn

Higher exposure than 98% of the 379 UK occupations we scored.

Read the score as
What a model can do unaided. A model plus workflow tools, closer to how work runs in 2026.

What this score means.

Most of the routine work in this role can already be done by a capable model. The role does not disappear, but its shape changes, and one person can credibly cover the work of several.

If you are in this role

Stop spending time on what a model can do well. Your value is judgement, relationships and the parts of the work that need you in the room. Redesign the workflow around that before someone else does it for you.

Almost every routine task in this role is within reach of today's models. Roles like this are being rebuilt, usually not by disappearing but by one person using AI to produce what several did.

If you are in this role

Be the person doing the rebuilding. People who learn to run this kind of work with AI become very valuable, and the ones who wait tend to have the change made for them.

Two clerks at the US Census Bureau punching population cards on a pantograph and a keypunch, with filing boxes behind them

Where the work usually starts.

These are the most important tasks a language model can already handle directly. In a typical engagement the first wins come from building workflows around them, so they stop eating the team's time.

  1. Design and create special programs for tasks such as statistical analysis and data entry and cleaning.

    O*NET importance 4.5 of 5 A model can do this directly

  2. Provide assistance with the preparation of project-related reports, manuscripts, and presentations.

    O*NET importance 4.2 of 5 A model can do this directly

  3. Perform descriptive and multivariate statistical analyses of data, using computer software.

    O*NET importance 4.1 of 5 A model can do this directly

These are the most important tasks AI can already handle when paired with the right tools and context. The first wins usually come from building workflows around them. That is the difference between a model that could do the job and one that actually does it inside your business.

  1. Design and create special programs for tasks such as statistical analysis and data entry and cleaning.

    O*NET importance 4.5 of 5 AI can do this with workflow tools

  2. Administer standardized tests to research subjects, or interview them to collect research data.

    O*NET importance 4.4 of 5 AI can do this with workflow tools

  3. Provide assistance with the preparation of project-related reports, manuscripts, and presentations.

    O*NET importance 4.2 of 5 AI can do this with workflow tools

Every role has three or four places like these. Finding them takes an hour. Turning them into a workflow your team actually uses takes a few days.

Every task in the role.

Every O*NET task for this occupation, split by what AI can do unaided and what still needs a person. Importance is O*NET's 1 to 5 rating of how central each task is.

What AI can already do

9 of 22 tasks, unaided

  1. Design and create special programs for tasks such as statistical analysis and data entry and cleaning.

    Importance 4.5 of 5

  2. Provide assistance with the preparation of project-related reports, manuscripts, and presentations.

    Importance 4.2 of 5

  3. Perform descriptive and multivariate statistical analyses of data, using computer software.

    Importance 4.1 of 5

  4. Verify the accuracy and validity of data entered in databases, correcting any errors.

    Importance 4.0 of 5

  5. Edit and submit protocols and other required research documentation.

    Importance 3.9 of 5

  6. Perform data entry and other clerical work as required for project completion.

    Importance 3.9 of 5

  7. Code data in preparation for computer entry.

    Importance 3.9 of 5

  8. Track laboratory supplies and expenses such as participant reimbursement.

    Importance 3.8 of 5

  9. Provide assistance in the design of survey instruments such as questionnaires.

    Importance 3.5 of 5

What still needs a person

13 of 22 tasks

  1. Obtain informed consent of research subjects or their guardians.

    Importance 4.5 of 5

  2. Administer standardized tests to research subjects, or interview them to collect research data.

    Importance 4.4 of 5

  3. Prepare tables, graphs, fact sheets, and written reports summarizing research results.

    Importance 4.2 of 5

  4. Recruit and schedule research participants.

    Importance 4.2 of 5

  5. Screen potential subjects to determine their suitability as study participants.

    Importance 4.2 of 5

  6. Track research participants, and perform any necessary follow-up tasks.

    Importance 4.1 of 5

  7. Develop and implement research quality control procedures.

    Importance 4.0 of 5

  8. Prepare, manipulate, and manage extensive databases.

    Importance 3.9 of 5

  9. Conduct internet-based and library research.

    Importance 3.8 of 5

  10. Present research findings to groups of people.

    Importance 3.5 of 5

  11. Supervise the work of survey interviewers.

    Importance 3.4 of 5

  12. Perform needs assessments or consult with clients to determine the types of research and information required.

    Importance 3.4 of 5

  13. Allocate and manage laboratory space and resources.

    Importance 3.0 of 5

What AI can already do

20 of 22 tasks, with tools

  1. Design and create special programs for tasks such as statistical analysis and data entry and cleaning.

    Importance 4.5 of 5

  2. Administer standardized tests to research subjects, or interview them to collect research data.

    Importance 4.4 of 5

  3. Provide assistance with the preparation of project-related reports, manuscripts, and presentations.

    Importance 4.2 of 5

  4. Prepare tables, graphs, fact sheets, and written reports summarizing research results.

    Importance 4.2 of 5

  5. Recruit and schedule research participants.

    Importance 4.2 of 5

  6. Screen potential subjects to determine their suitability as study participants.

    Importance 4.2 of 5

  7. Perform descriptive and multivariate statistical analyses of data, using computer software.

    Importance 4.1 of 5

  8. Track research participants, and perform any necessary follow-up tasks.

    Importance 4.1 of 5

  9. Verify the accuracy and validity of data entered in databases, correcting any errors.

    Importance 4.0 of 5

  10. Develop and implement research quality control procedures.

    Importance 4.0 of 5

  11. Prepare, manipulate, and manage extensive databases.

    Importance 3.9 of 5

  12. Edit and submit protocols and other required research documentation.

    Importance 3.9 of 5

  13. Perform data entry and other clerical work as required for project completion.

    Importance 3.9 of 5

  14. Code data in preparation for computer entry.

    Importance 3.9 of 5

  15. Track laboratory supplies and expenses such as participant reimbursement.

    Importance 3.8 of 5

  16. Conduct internet-based and library research.

    Importance 3.8 of 5

  17. Provide assistance in the design of survey instruments such as questionnaires.

    Importance 3.5 of 5

  18. Present research findings to groups of people.

    Importance 3.5 of 5

  19. Supervise the work of survey interviewers.

    Importance 3.4 of 5

  20. Perform needs assessments or consult with clients to determine the types of research and information required.

    Importance 3.4 of 5

What still needs a person

2 of 22 tasks

  1. Obtain informed consent of research subjects or their guardians.

    Importance 4.5 of 5

  2. Allocate and manage laboratory space and resources.

    Importance 3.0 of 5

Stay on top of this.

An email when there is something worth reading: what is real in AI at work, what is noise, and what operators are doing about it.

Free. Unsubscribe any time.

What it means for your sector.

All guides

This role's score comes from Eloundou et al's 2023 task labels, weighted by O*NET importance within each O*NET-SOC code, then bridged to UK SOC 2020 through ISCO-08 (the ONS coding index) and US SOC 2010 (the BLS crosswalk). Employment and median pay come from ONS ASHE Table 14.7a, 2025 provisional, which covers employees only.

Pictures

An email when there’s something worth reading.

What I’m building, what’s working, and what it cost. Unsubscribe in one click.

Free. No spam.