Data Scientists
Develop and implement a set of techniques or analytics applications to transform raw data into meaningful information using data-oriented programming languages and visualization software. Apply data mining, data modeling, natural language processing, and machine learning to extract and analyze information from large structured and unstructured datasets. Visualize, interpret, and report data findings. May create dynamic data reports.
On a phone, press and hold a picture to save it to Photos.
opens your phone's share sheet, where you can save both to Photos
What the data says: AI could halve the time on 15 of its 16 tasks, yet jobs are projected to grow 34.6% by 2035. In the most exposed occupations, the squeeze shows up first among early-career workers. How to think about this →
About this projection
BLS builds AI into its projections where the research supports it: “BLS generally applies adjustments based on this research conservatively, where there is convincing evidence for a change.”
It also notes a limit: “BLS projection methods are not designed to capture extremely rapid technological change.”
On its AI exposure groups, BLS says exposure “does not imply job loss.”
BLS figures cover the broader occupation “Data scientists”, which includes this one and others.
Recent conditions
- In 2024, 2.9% of people in this work were unemployed, against 3.8% across all occupations Census ACS 2024, all workers in the Census group “Other mathematical science occupations”
- From Sep 2025 to Aug 2026, 7.9 hires for every 10 job openings in professional and business services, which employs about 44% of people in this work, against 7.3 across all industries BLS JOLTS
One box per opening, filled for each hire; hires beyond the openings are added in a lighter shade. Higher can mean openings get filled, or that people leave often; lower, that jobs go unfilled, or hiring is slow.7.9 this industry7.3 all industries
Median pay by state BLS OEWS May 2025, state estimates
Tap a state for its figures.
Median pay in every state
| State | Median | Middle half earn | Jobs |
|---|---|---|---|
| Alabama | $103,120 | $74,140 to $127,660 | 1,740 |
| Alaska | $84,110 | $57,350 to $113,040 | 230 |
| Arizona | $107,240 | $80,820 to $134,070 | 4,560 |
| Arkansas | $109,390 | $97,510 to $136,060 | 530 |
| California | $141,590 | $103,360 to $186,820 | 39,310 |
| Colorado | $117,400 | $96,300 to $152,650 | 6,280 |
| Connecticut | $126,340 | $94,480 to $163,210 | 1,760 |
| Delaware | not published | not published | 770 |
| District of Columbia | $126,490 | $105,320 to $174,650 | 2,680 |
| Florida | $115,820 | $84,070 to $140,210 | 10,240 |
| Georgia | $104,340 | $84,120 to $136,560 | 9,260 |
| Hawaii | $102,130 | $87,640 to $152,980 | 300 |
| Idaho | $89,380 | $80,940 to $136,940 | 1,350 |
| Illinois | $106,560 | $84,610 to $137,290 | 10,520 |
| Indiana | $91,470 | $72,670 to $127,510 | 3,580 |
| Iowa | $95,830 | $71,010 to $127,340 | 2,300 |
| Kansas | $98,130 | $75,970 to $131,390 | 490 |
| Kentucky | $91,760 | $75,620 to $121,850 | 1,540 |
| Louisiana | $78,760 | $63,560 to $110,130 | 1,300 |
| Maine | $90,880 | $77,340 to $115,170 | 1,210 |
| Maryland | $136,370 | $100,370 to $169,510 | 3,340 |
| Massachusetts | $131,750 | $101,940 to $166,180 | 9,420 |
| Michigan | $100,590 | $77,410 to $130,240 | 7,360 |
| Minnesota | $128,800 | $92,490 to $160,910 | 4,020 |
| Mississippi | $69,490 | $51,060 to $99,030 | 440 |
| Missouri | $97,090 | $74,610 to $127,680 | 4,730 |
| Montana | $100,490 | $70,990 to $127,840 | 220 |
| Nebraska | $98,230 | $80,900 to $109,860 | 1,980 |
| Nevada | $98,540 | $76,190 to $129,990 | 1,680 |
| New Hampshire | $100,620 | $82,750 to $128,540 | 1,100 |
| New Jersey | $135,280 | $101,970 to $167,330 | 6,430 |
| New Mexico | $95,850 | $64,250 to $107,540 | 560 |
| New York | $130,460 | $94,880 to $171,580 | 23,970 |
| North Carolina | $119,090 | $82,440 to $147,540 | 11,430 |
| North Dakota | $81,900 | $64,270 to $100,110 | 210 |
| Ohio | $102,630 | $79,510 to $130,920 | 6,100 |
| Oklahoma | $86,310 | $68,580 to $109,770 | 1,900 |
| Oregon | $125,990 | $96,770 to $163,710 | 2,140 |
| Pennsylvania | $106,850 | $79,220 to $129,160 | 13,810 |
| Puerto Rico | $80,350 | $66,750 to $97,320 | 100 |
| Rhode Island | $102,440 | $93,090 to $130,150 | 1,160 |
| South Carolina | $91,990 | $78,550 to $123,880 | 3,750 |
| South Dakota | $96,150 | $83,810 to $123,260 | 220 |
| Tennessee | $100,330 | $77,770 to $131,930 | 2,790 |
| Texas | $122,090 | $83,880 to $146,030 | 25,860 |
| Utah | $108,090 | $78,720 to $146,160 | 4,360 |
| Vermont | $127,070 | $98,450 to $143,170 | 200 |
| Virginia | $126,430 | $97,390 to $165,610 | 8,920 |
| Washington | $163,350 | $105,240 to $208,320 | 9,600 |
| West Virginia | $85,090 | $69,790 to $112,770 | 390 |
| Wisconsin | $106,680 | $88,060 to $133,860 | 4,370 |
Nationally: $120,230. Shades run from the lowest state median to the highest; a territory below that range takes the lightest. How to read this
How AI touches the work
What AI could speed up, task by task, and where people already use it.
AI could halve the time on 15 of its 16 tasks
| AI rating | Its main tasks | Share of Claude conversations May 2026 |
|---|---|---|
| AI + software | Analyze, manipulate, or process large sets of data using statistical software. | 0.03% |
| AI + software | Create graphs, charts, or other visualizations to convey the results of data analysis using specialized software. | 0.36% |
| AI + software | Test, validate, and reformulate models to ensure accurate prediction of outcomes of interest. | 0.01% |
| AI + software | Deliver oral or written presentations of the results of mathematical modeling and data analysis to management or other end users. | under 0.01% |
| AI alone | Recommend data-driven solutions to key stakeholders. | 0.01% |
| Not much | Compare models using statistical performance metrics, such as loss functions or proportion of explained variance. | under 0.01% |
See all 16 tasks
| AI + software | Clean and manipulate raw data using statistical software. | 0.02% |
| AI + software | Identify business problems or management objectives that can be addressed through data analysis. | Not published |
| AI + software | Apply feature selection algorithms to models predicting outcomes of interest, such as sales, attrition, and healthcare use. | Not published |
| AI + software | Identify solutions to business problems, such as budgeting, staffing, and marketing decisions, using the results of data analysis. | 0.02% |
| AI + software | Identify relationships and trends or any factors that could affect the results of research. | 0.01% |
| AI alone | Write new functions or applications in programming languages to conduct analyses. | 0.12% |
| AI + software | Read scientific articles, conference papers, or other sources of research to identify emerging analytic trends and technologies. | Not published |
| AI alone | Apply sampling techniques to determine groups to be surveyed or use complete enumeration methods. | Not published |
| AI + software | Design surveys, opinion polls, or other instruments to collect data. | Not published |
| AI alone | Propose solutions in engineering, the sciences, and other fields using mathematical theories and techniques. | 0.04% |
Ordered by how important O*NET's respondents rate each task (1 to 5) times the share who say it is part of their job. The ordering is ours. 11 of the 16 have a published share of Claude conversations; “not published” means too few conversations for Anthropic to publish a figure.
Where people already use AI
The tasks above say what AI could speed up. This shows what people are using it for. Its 16 tasks fall under 13 work activities that many occupations share, and Microsoft counted how often each one came up in 200,000 conversations with its Copilot assistant in 2024.
- #4 Maintain current knowledge in area of expertise
- #14 Create visual designs or displays
- #20 Present research or technical information
- and 5 more, in the table below
| Work activity | This occupation's tasks under it | People asked AI for help in 10,000 conversations |
AI did it for them in 10,000 conversations |
|---|---|---|---|
| Analyze scientific or applied data using mathematical principles | |||
| Present research or technical information | |||
| Analyze data to improve operations | |||
| Maintain current knowledge in area of expertise | |||
| Create visual designs or displays | |||
| Program computer systems or production equipment | |||
| Process digital or online data | |||
| Determine operational methods or procedures |
Show all 13 activities
| Select materials or equipment for operations or projects | |||
| Analyze business or financial data | |||
| Develop operational or technical procedures or standards | |||
| Advise others on business or operational matters | |||
| Develop scientific or mathematical theories or models |
How people in this work use Claude
What people in this work made with Claude % of conversations
- A chart or visualization25%
- A presentation or slides17%
- A script or snippet14%
- A document or report10%
- Data or a spreadsheet9%
How they worked with Claude
56% of these conversations were about work. Anthropic Economic Index, May 2026.
How to read this
Work activities come from O*NET, which links every task to the broader activities it involves. Microsoft's figures are for the activity across everyone who used Copilot, not for data scientists specifically. Copilot is a general chat assistant; dedicated coding tools are not included, so programming help is under-counted here. Activities are listed by how many of this occupation's tasks fall under them.
The task ratings say whether AI could halve the time a task takes: on its own (“AI alone”), or with software built on it (“AI + software”). They were made by people in 2023 (Eloundou et al.). Claude's share is the task's share of all Claude.ai conversations in May 2026; Anthropic doesn't publish a task with too few conversations. None of these add up to the BLS exposure group, which combines five research measures.
More in the Guide: task ratings · task order · work activities · Copilot figures · Claude
High exposure doesn't mean fewer jobs
BLS projects most occupations in its two most exposed groups to grow: 141 of 206 rated Very high and 165 of 206 rated High. Each dot is one of the 831 occupations BLS rates.
Career moves into and out of this occupation
From moves between occupations found in online resumes, 2002–2016. Career changes only: a first job after school isn't a move.
Where people moved in from
of 675 moves in
- 1 Janitors and Cleaners, Except Maids and Housekeeping Cleaners Low +2.2% 23%
- 2 Operations Research Analysts Very high +11.9% 13%
- 3 Software Developers, Applications (2010 group) an older group BLS has since split; no exposure group today 8%
- 4 Information Security Analysts Very high +21.0% 7%
- 5 Mechanical Engineers Very high +11.2% 6%
The resume data uses 2010 occupation codes. Moves into this occupation are counted as moves into “Mathematical Science Occupations, All Other”, which today also includes the occupation still called “Mathematical Science Occupations, All Other”, so these figures cover it too. Resumes over-represent younger, white-collar job-seekers and moves into management. Schubert, Stansbury and Taska (resume data).
Education and training for the occupation
Census survey (American Community Survey)
- 1% less than high school
- 4% high school
- 7% some college
- 4% associate's
- 39% bachelor's
- 35% master's
- 2% professional degree
- 10% doctorate
Census ACS 2024, all workers in the Census group “Other mathematical science occupations” · 4,447 people surveyed.
Median pay by highest degree full-time, year-round · Census ACS 2024, all workers in the Census group “Other mathematical science occupations”
35 programs, such as Artificial Intelligence, Information Technology and Computer Science.
See all
- Artificial Intelligence
- Information Technology
- Computer Science
- Data Modeling/Warehousing and Database Administration
- Modeling, Virtual Environments and Simulation
- Bioinformatics
- Computational Biology
- Biomathematics, Bioinformatics, and Computational Biology, Other
- Mathematics, General
- Applied Mathematics, General
- Computational Mathematics
- Computational and Applied Mathematics
- Financial Mathematics
- Mathematical Biology
- Applied Mathematics, Other
- Statistics, General
- Mathematical Statistics and Probability
- Mathematics and Statistics
- Applied Statistics, General
- Mathematics and Statistics, Other
- Mathematics and Computer Science
- Computational Science
- Economics and Computer Science
- Linguistics and Computer Science
- Data Science, General
- Data Science, Other
- Data Analytics, General
- Business Analytics
- Data Visualization
- Financial Analytics
- Data Analytics, Other
- Cheminformatics/Chemistry Informatics
- Econometrics and Quantitative Economics
- Management Science
- Business Statistics
Jobs that require a credential BLS Occupational Requirements Survey 2023
BLS publishes no credential figures for this occupation.
Look up certifications, training and licenses U.S. Department of Labor
- Certification Finder ↗ certifications CareerOneStop lists for this occupation
- Training Finder ↗ schools and programs, nationwide; narrow the place there
- License Finder ↗ licenses by state
- Apprenticeship Finder ↗ DOL lists 2 registered apprenticeships in this occupation (Data Scientist; Machine Learning Data Curator). Apprenticeship.gov can't be linked to one occupation, so search for “Data Scientists” there.
These open in a new tab and send only the occupation's public code. More in the Guide
BLS AI exposure categories (Aug 2026) · BLS Employment Projections 2025–35 · O*NET 31.0 · Task ratings: Eloundou et al., Science (2024) · Microsoft "Working with AI" v1.1 · Anthropic Economic Index (2026-06-26) · Schubert, Stansbury and Taska · NCES CIP–SOC · Census ACS 2024 · BLS JOLTS · Brynjolfsson, Chandar and Chen, Stanford (Aug 2026)
See it online, and explore any career: https://www.pathmosaic.com/card/15-2051.00