Data scientist salary: what data scientists earn and what raises pay

Data scientists in the US earned a median of $120,230 a year in May 2025, according to the Bureau of Labor Statistics. This guide shows you where that figure comes from, how wide the range really is, which industries pay above it, and how to use the official numbers to check whether a data science offer is fair.

Half of all data scientists earned more than that median and half earned less. It is a midpoint, not a promise, and the spread around it is wide.

Quick reference: data scientist pay, May 2025

All figures below are for the BLS occupation "Data Scientists" (SOC code 15-2051) from the Occupational Employment and Wage Statistics (OEWS) survey, labeled May 2025, which is the newest set BLS has published.

A quick word on those terms. The median is the middle paycheck if you lined up every data scientist from lowest to highest paid. The 10th percentile is the point where only 1 in 10 earn less, and the 90th percentile is where only 1 in 10 earn more. The mean is the plain average, which gets pulled up by a small group of very high earners. That is why the average ($126,800) sits above the median ($120,230).

What data scientists earn nationally

The $120,230 median puts data scientists well above the typical US worker. BLS lists the median for all occupations at $50,980 in May 2025, and the median for the wider group of mathematical science occupations at $107,570.

On an hourly basis, the median works out to $57.80, and the average to $60.96. Most data scientists work full time, so the annual figure is the one to use when you are comparing salaried offers.

Two counts of jobs float around, and they measure different things:

• 262,440 is the OEWS count of wage and salary jobs in May 2025. It is the base for every pay figure in this article.
• 275,600 is the BLS projections figure for 2025, used in the Handbook's job outlook. BLS projects employment to grow 35 percent from 2025 to 2035, with about 24,800 openings a year on average.

Strong projected demand does not guarantee a raise for you, but it is a reasonable point to bring up when you explain why your skills are hard to replace.

What the OEWS number includes, and what it leaves out

This matters more for data scientists than for most jobs, because so much tech pay arrives outside the base salary.

The OEWS technical note defines its wages as straight-time, gross pay. Here is how that plays out:

Stock is not on the included list. Restricted stock units, stock options and employee stock purchase discounts are not part of the straight-time wage OEWS measures. So if you work at a public tech company where equity is a large slice of pay, the BLS figure will understate your total compensation, sometimes by a lot. BLS does not publish equity data for this occupation, so there is no official number to tell you how big that gap is.

There is also a timing lag. The May 2025 estimates pool survey responses from six panels collected over three years, from November 2022 through May 2025. Treat the figures as a solid baseline rather than a reading of this month's job market.

The percentile range and who sits at the top

The gap between the 10th and 90th percentiles is wide: less than $67,240 at the bottom and more than $199,130 at the top. The top-tenth threshold is about three times the bottom-tenth one, inside a single job title.

Who tends to land near each end? The OEWS tables do not break pay down by experience, so this part is about what the data can and cannot show.

• Near the 10th percentile: expect entry-level roles, jobs in lower-paying industries and regions, and positions where "data scientist" covers work closer to reporting and dashboards.
• Around the median: experienced individual contributors in mainstream industries such as insurance, banking and corporate headquarters.
• Above the 75th percentile and toward the 90th: senior and staff-level people, roles that build and ship machine learning systems, and employers in the industries listed in the next section.

The Handbook only shows the 10th, 50th and 90th percentiles. The 25th and 75th percentiles for this occupation sit in the OEWS profile tool. If you want them, pull them from the May 2025 OEWS occupation profiles under 15-2051.

Pay by industry

Industry is one of the biggest levers on data science pay. The Handbook lists median annual wages in May 2025 for the five industries that employ the most data scientists:

A few things to notice:

• The top group is not traditional newspapers. Under the current industry classification, publishing includes software publishers, and the content providers side takes in online media platforms. Tech product companies show up here.
• Insurance is a big employer of data scientists but pays below the national median for the role. If you are coming from insurance, moving into banking or tech services is one of the clearest routes to a higher number.
• Consulting-style firms (computer systems design) pay above the median, but check how much of your time would be billable and whether bonuses depend on utilization.

These are medians. A data scientist at a large insurer can still out-earn one at a small software firm.

Top-paying states and metro areas

Location changes pay a lot for data scientists, but here is the honest part: state and metro rankings should come straight from the OEWS tables, and those tables for May 2025 sit in the BLS profile tool rather than in the summary pages and Handbook figures used above. This article does not repeat third-party "best cities" lists, because many mix survey years or use job-ad data.

Here is how to pull the official figures for your area in a few minutes:

1. Open the May 2025 OEWS occupation profiles and choose 15-2051 Data Scientists.
2. Look at the state and metro tables in the profile. BLS "top paying" lists for an occupation have traditionally been sorted by annual mean wage, so note both the mean and the median.
3. For your own area, check the May 2025 metro area estimates or the state estimates and find the data scientist row.
4. Write down the median, the 10th and 90th percentiles and the employment count.

When you read those tables, keep three things in mind:

• Averages can mislead. A mean wage can be pushed up by a few very high salaries. Check the median as well before you decide an area pays more.
• Small areas are noisy. A metro with only a few hundred data scientists can swing year to year. Where BLS judges an estimate unreliable, it suppresses it rather than printing a shaky number.
• High pay often comes with high costs. The Bureau of Economic Analysis publishes Regional Price Parities, which compare price levels across states and metro areas. Use them to adjust a salary before you compare two cities. A higher-paying metro can still leave you with less after rent.

If you are weighing a remote offer, ask which location the employer uses to set pay. Some set ranges by where you live, others by where the team sits.

How data scientist pay compares with neighboring roles

Job title can shift pay as much as skill does, because employers report workers under the occupation that best matches their duties. The table below uses the May 2025 OEWS national figures.

For annual medians, the Handbook lists $88,940 for operations research analysts and $140,300 for computer and information research scientists in May 2025. It combines mathematicians and statisticians into one profile with a median of $105,720.

What this tells you:

• Statistician vs. data scientist: similar math, lower median pay. If your work involves building production models, coding pipelines or machine learning, the data scientist title (and its pay band) may fit your job better.
• Operations research analysts earn noticeably less at the median, even though the work overlaps on optimization and modeling.
• Software developers and computer and information research scientists sit higher. Engineering-heavy machine learning work often lands here.

Why "machine learning engineer" is not in the BLS data

OEWS uses the 2018 Standard Occupational Classification, about 830 occupations in the May 2025 estimates, and none of them is called machine learning engineer, AI engineer or MLOps engineer. Employers who fill out the survey decide which occupation best matches each worker's duties.

Depending on the job, an ML engineer might be reported as a data scientist, a software developer, a computer and information research scientist, or under "computer occupations, all other." So:

• Do not compare an ML engineer offer only against the data scientist median. Check the software developer and research scientist rows too.
• If your job is mostly modeling and analysis, the data scientist figures are the closer match, whatever your title says.
• Titles like "analytics engineer" or "decision scientist" have the same problem. Match on duties, not on the job ad's label.

What moves pay up: specialization, education and seniority

BLS says data scientists typically need at least a bachelor's degree in mathematics, statistics, computer science or a related field, and that some employers require or prefer a master's or doctorate. The Handbook does not list a license or on-the-job training requirement for entry.

The OEWS data does not split pay by degree or years of experience, so the following is a guide to the levers rather than a promise of dollar gains:

• Engineering depth. The Handbook notes that some data scientists with coding or engineering backgrounds build machine learning algorithms and systems. That work overlaps with the higher-paid software developer and research scientist occupations shown above.
• Industry knowledge. BLS points out that some employers want industry-related experience, giving finance as an example. Domain knowledge in banking, insurance pricing or healthcare can make you harder to replace.
• Advanced degrees. A master's or PhD can open roles that list one as a requirement, especially research-focused teams. The research scientist occupation lists a master's degree as typical entry-level education.
• Seniority and scope. Moving from individual contributor to lead, staff or principal level usually means owning bigger problems. Moving into management is a separate path: computer and information systems managers had a mean annual wage of $192,160 in May 2025.
• Communication. BLS lists communication skills among the key qualities. People who can turn a model into a business decision tend to get bigger projects, and bigger projects make the strongest case at review time.

A typical progression looks like this: an analyst or junior data scientist role, then a mid-level data scientist role where you own projects, then a senior or staff role, and from there either a deeper technical track or a management track. Titles and levels vary a lot between employers, so compare scope, not labels.

How to benchmark a data science offer

Here is a step-by-step check you can run before you reply to an offer.

1. Find your occupation by duties. Data scientist, software developer, research scientist or statistician? Use the row that best matches the work.
2. Place the base salary on the national range. Below $67,240 puts it in the bottom tenth for data scientists nationally. Around $120,230 is the middle. Above $199,130 is top-tenth territory.
3. Adjust for place. Pull the state or metro median from OEWS, then look at cost of living using BEA Regional Price Parities.
4. Adjust for industry. If the employer is in software, content platforms or banking, the national median understates what peers there earn.
5. Add what OEWS leaves out. Ask for the annual bonus target, the equity grant (number of shares or units, vesting schedule and the value the company uses), sign-on bonus and benefits. Compare total compensation separately from base.
6. Check the posted range. In places with pay transparency rules, the job ad or the employer may have to share a salary range. Use the range to see where your offer falls.

Example (hypothetical): you are offered a base salary right at the $120,230 national median for a senior role at a software company. The industry table shows a median of $142,240 for the publishing and content group, which includes software publishers, and seniority usually puts you above the median, not on it. That is a reasonable basis for asking for more base pay.

Scripts you can use

At offer stage:

"Thanks, I'm excited about this role. Looking at BLS data for data scientists, the national median was $120,230 in May 2025, and this industry pays above that. Given that I'll be building production models, I was hoping for a base closer to the top of your range. Is there flexibility there?"

When equity is part of the deal:

"Could you walk me through the equity: how many units, the vesting schedule, and the share price you used to value it? I want to compare total compensation, not just base."

At review time:

"Over the past year I shipped [project] and it [result]. My role now overlaps with what BLS classifies as software development, where pay runs higher. I'd like to talk about moving my pay toward that level."

Keep the ask specific and tied to your work. Never claim a number you cannot back up, and be ready for the employer to say the range is fixed. If base pay will not move, ask about a sign-on bonus, a larger equity grant or a review date in six months.

Pay transparency is on your side

A growing number of states and cities require employers to include salary ranges in job postings or to share them on request. Rules differ by location and by employer size, so check your state labor department's website for what applies to you. Posted ranges are one of the few sources that capture current pay at a named employer, which makes them a useful complement to the BLS baseline.

What to do next

• Pull the May 2025 OEWS figures for 15-2051 in your state or metro, and for one neighboring occupation that matches your duties.
• Write down your offer or current pay as base, bonus target and annual equity value, separately.
• Pick the one or two strongest reasons you belong above the median (industry, scope, specialist skills) and build your ask around those.
• Revisit the numbers each spring, when BLS usually releases a new OEWS year, and before every review.

For more pay breakdowns by job, see salaries by role, and for help with the conversation itself, see negotiation.