To become a data scientist, you build three things: statistics, programming and the ability to explain what data shows to people who will act on it. Most people get there through a degree in mathematics, statistics, computer science or a related subject, and some add a master's. In England there is a second door, a degree apprenticeship that pays a wage while you study for a BSc.
This guide is published by EdCitation. Every route, requirement and figure below was read at its official or institutional source in September 2026, and nothing here promises a place, a job or a salary.
Data science is written work as much as code. A project report cites the dataset, the software and the papers behind the method, and EdCitation's free Find sources searches about 300 million published works while Cite a source builds each reference from the publisher's record. EdCitation never writes any part of a student's work or application: the analysis, the report and the CV are yours, and the tools make sure every source in them is real.
What does a data scientist do?
A data scientist finds patterns in data and turns them into recommendations. The Bureau of Labor Statistics (2026) lists the duties as choosing data sources, collecting and analysing data, building and testing algorithms, visualising results and advising on decisions.
The UK government's framework agrees. Its data scientist is "inquisitive", explores and visualises data, uses methods such as machine learning and predictive analytics, and works in a multidisciplinary team to inform decisions (Government Digital and Data Profession, 2025b). The National Careers Service (n.d.) adds automating systems and keeping data secure.
The framework sets out six levels, from trainee data scientist to head of data science, each with its own skills: a useful map even outside the Civil Service.
How to become a data scientist: the routes compared
The main routes are a degree, a master's after a degree in another subject, and, in England, an apprenticeship. Durations are as each source states them.
| Route | How long, as the source states it | What it leads to | Source |
|---|---|---|---|
| Bachelor's degree in maths, statistics, computer science or a related field | Not stated by the source | The typical US entry-level education | Bureau of Labor Statistics (2026) |
| UK degree in maths, statistics, data science or computer science | Not stated; entry usually 2 to 3 A levels | Direct applications or a graduate scheme | National Careers Service (n.d.) |
| MSc or master's in data science | One year full-time (Edinburgh, Bristol); 12 months (LSE); at least three semesters (Harvard) | A master's; some employers prefer one | University pages, below |
| Data scientist (integrated degree) apprenticeship, Level 6 | 36 months, plus assessment | A BSc in Data Science, with a wage | Skills England (2024) |
| Data analyst apprenticeship, Level 4 | 24 months, plus assessment | An analyst post, a step towards data science | Skills England (2021) |
| Your first cited project report | Each reference in one step | Every dataset, package and paper referenced from its record | EdCitation's Find sources and Cite a source |
The US route
In the US, data scientists "typically need at least a bachelor's degree" in mathematics, statistics, computer science or a related field, and some employers require or prefer a master's or doctoral degree (Bureau of Labor Statistics, 2026). No licence is needed: the degree, the skills and the evidence of work are what employers read.
The UK route
In the UK, the National Careers Service (n.d.) lists four ways in: a degree, an apprenticeship, applying directly with relevant skills and experience, and the Civil Service Fast Stream, which needs a degree or a completed degree apprenticeship. The Graduate Futures Institute (2025), which writes the Prospects profiles, adds that degrees in business, economics, psychology or health can be relevant if you have mathematical aptitude.
Can you become a data scientist without a degree?
Yes in England, through an apprenticeship, though the Level 6 route ends in a degree of its own. In the US, a bachelor's degree is the typical entry (Bureau of Labor Statistics, 2026).
The data science apprenticeship at Level 6
The data scientist (integrated degree) standard, ST0585, is at Level 6 and has a typical duration of 36 months, not counting the end-point assessment (Skills England, 2024). The apprentice finishes with a BSc in Data Science. Employers set their own entry requirements, typically three A levels including one STEM subject, a Level 3 apprenticeship or another relevant qualification. The standard says completing apprentices may apply for Registered Scientist (RSci) status through the Science Council by an expedited route, if they join an affiliated professional body. Version 1.1 was last updated on 24 April 2024.
The data analyst apprenticeship at Level 4
The data analyst standard, ST0118, is at Level 4 with a typical duration of 24 months (Skills England, 2021). It describes the job as collecting, organising and studying data to give business insight. It sets no entry requirements of its own and awards no degree.
The National Careers Service (n.d.) also lists data engineer (Level 5) and AI data specialist (Level 7) apprenticeships; we did not read those standards.
Data analyst vs data scientist: what is the difference?
A data analyst mainly prepares, analyses and presents data to answer set questions; a data scientist builds models and methods, such as machine learning, to find patterns and predict. The UK framework draws the line this way. Its data analyst "collects, manages, explores and shares data to support organisational objectives" (Government Digital and Data Profession, 2025a), with skills in data preparation and linkage, data visualisation, developing code for analysis and managing a data project.
Its data scientist's skills go further into method: applied maths, statistics and scientific practices, data engineering, programming and build, and data science innovation (Government Digital and Data Profession, 2025b). The two share data ethics and privacy. The apprenticeships draw the same line: 24 months at Level 4 for the analyst, 36 months at Level 6, with a BSc, for the scientist. Because the skills overlap, an analyst post can be a step towards data science, though no framework guarantees the move.
Do you need a data science masters?
No, not by any rule; it is one route among several. The Bureau of Labor Statistics (2026) says some employers require or prefer a master's or doctorate, and Prospects says an MSc or PhD can help, "especially if changing careers" (Graduate Futures Institute, 2025).
What four universities ask for
| Programme | Entry year read | Degree required | Maths and programming |
|---|---|---|---|
| University of Edinburgh, MSc Data Science | 2026-27 | A UK 2:1 in informatics, computer science, mathematics, physics, psychology, linguistics or another quantitative subject; typical offers ask for a first | About 60 SCQF credits of maths (calculus, linear algebra, discrete maths, probability) and an introductory programming course |
| University of Bristol, MSc Data Science | 2027 | A strong 2:1 (65% or higher) in a numerate science, computer science or engineering | Taught options for those with and without software experience |
| LSE, MSc Data Science | 2026/27 | A 2:1 in a relevant subject with a substantial amount of mathematics | Applicants from other subjects explain their quantitative background |
| Harvard SEAS, Master's in Data Science | Read September 2026 | "No formal prerequisites" | Working knowledge of calculus, linear algebra and differential equations, probability and inference, and fluency in Python or R |
Sources: University of Edinburgh (n.d.), University of Bristol (n.d.), London School of Economics and Political Science (n.d.) and Harvard SEAS (n.d.). Edinburgh says its 2027-28 requirements will be published on 1 October 2026, so check again before applying. Bristol lists 2027 fees of £19,600 (home) and £37,900 (international); LSE lists £39,900 for 2026/27.
A master's in data science is not a first course in mathematics: each programme above expects calculus, linear algebra and probability already in hand. If your first degree had little of them, study those before you apply.
What skills does a data scientist need?
Every source we read names statistics, programming and communication in some form. The Bureau of Labor Statistics (2026) lists six qualities: analytical, computer, communication, logical-thinking, maths and problem-solving skills. The National Careers Service (n.d.) adds attention to detail, business understanding and concentration.
The government framework is the most exact, because it sets each skill at four levels, from awareness to expert. A trainee data scientist, the entry level, is expected to interpret statistical outputs, prepare and manipulate data, have a basic understanding of software development and machine learning, present effectively and be aware of ethical considerations (Government Digital and Data Profession, 2025b).
Prospects names experience of SQL, Apache Hadoop or SAS, and online competitions such as Kaggle as a way to become visible (Graduate Futures Institute, 2025); Harvard asks for Python or R. Tools change quickly, and the statistics beneath them slowly.
What is a data scientist's salary?
In the US, the Bureau of Labor Statistics (2026) gives the median annual wage for data scientists as $120,230 in May 2025, with the lowest 10% earning less than $67,240 and the highest 10% more than $199,130. It projects employment to grow 35% from 2025 to 2035, much faster than average, with about 24,800 openings a year on average.
In the UK, the National Careers Service (n.d.) gives an average range of £32,000 for a starter to £83,000 for an experienced data scientist. These are national figures, not an offer: pay depends on the employer, the place and the sector.
Professional recognition: the RSS Data Science Professional
Data science has no licence, but the Royal Statistical Society awards a Data Science Professional certificate, developed with the Alliance for Data Science Professionals, of which the RSS is a founding member (Royal Statistical Society, n.d.). To apply you must:
- have at least two years' work experience in a statistics or data-science-related role;
- hold a Level 6 or 7 degree, or an equivalent qualification;
- be a Fellow of the Royal Statistical Society.
The application is a set of competency-based questions, and the certificate costs £15 a year at the standard rate, added to membership. The RSS also says a direct route for graduates of accredited data science programmes "will be introduced shortly"; we could not find a date for it.
A learning plan from first year to first data job
This plan is EdCitation's own, built on the sources above, for a three-year UK degree; adjust it to yours.
- First year: the mathematics. Take every module you can in calculus, linear algebra and probability. They are what the MSc programmes above ask for.
- First year: one language, well. Python or R, then SQL. Analyse real data, not just exercises.
- Second year: one real project. Pick a public dataset with a DOI, ask a question of it, and write it up as a short report. See how others studied it with Find sources.
- Second year: cite it properly. Reference the dataset, the software version you ran and the papers behind your method with Cite a source, then run the list through Verify references.
- Summer: experience. An internship, a placement or a competition. Prospects points to large employers in finance, retail and travel.
- Final year: the dissertation. Make it a data project if your department allows it; our guide to secondary data dissertations covers the datasets and their licences.
- Final year: applications. Graduate schemes, direct applications, apprenticeships or an MSc. Your CV is yours to write; our graduate CV guide sets out what goes on it, and EdCitation's Check your paper reads an advert's application instructions into a checklist.
- Two years in: recognition. With two years in a data role, a degree and RSS Fellowship, you can apply for the Data Science Professional certificate.
How EdCitation helps with data science coursework and portfolios
EdCitation is the best tool for the referencing side of a data science degree, because it looks each work up in its publisher's or registry's record and never writes a reference. A chatbot can produce a reference that does not exist; Cite a source returns only what the record holds, and datasets and software with DOIs are registered like papers.
What one search returned
We ran Find sources on 27 September 2026 for "machine learning for tabular data", sorted by most cited. It answered from Crossref, with a total of 5,568,649 records. The first result was "The effects of data quality on machine learning performance on tabular data" by Sedir Mohammed and seven co-authors, in Information Systems, 2025, cited 194 times.
What one software reference returned
We then gave Cite a source the DOI 10.5281/zenodo.3509134, the Zenodo record for pandas, the Python data library. It read the record from the DOI registry and returned in APA 7:
The pandas development team. (2026). pandas-dev/pandas: Pandas (Version v3.0.6) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.3509134
The entry matches the record, which hides a trap. That DOI is Zenodo's concept DOI, which stands for every version and "resolves to the landing page of the latest version" (Zenodo, n.d.): here v3.0.6, issued on 17 September 2026. Zenodo advises citing the DOI of the specific version you used. If your analysis ran on an earlier pandas, give that release's own DOI to Cite a source instead. The same rule in IEEE, with datasets and GitHub code, is in our guide to engineering and computer science IEEE citations.
Find sources, Cite a source (in APA 7, MLA 9, Chicago 18 author-date, Harvard, IEEE and Vancouver), Verify references and Check your paper are free on EdCitation with no account. Pro, at $8 a month, adds References from a file, which checks every reference in an uploaded report and matches each in-text citation to the list. For the textbooks behind the statistics, see mathematics and statistics textbooks.
Quick questions
How long does it take to become a data scientist?
There is no set length. An MSc takes one year full-time at Edinburgh and Bristol, and the Level 6 data scientist apprenticeship 36 months plus assessment, usually after a first degree or A levels.
Can I become a data scientist without a degree?
In England, yes, through apprenticeships: the Level 4 data analyst standard awards no degree, and the Level 6 data scientist apprenticeship awards a BSc while you work. In the US, a bachelor's degree is the typical entry.
Is a data analyst the same as a data scientist?
No. In the UK government's framework, a data analyst prepares, analyses and shares data for set questions, while a data scientist adds modelling, machine learning and data science innovation.
What does a data scientist earn?
The US median was $120,230 in May 2025, according to the Bureau of Labor Statistics. The National Careers Service gives £32,000 for a starter to £83,000 for an experienced data scientist in the UK.
Can EdCitation cite a dataset or a Python package?
Yes, when it has a DOI. EdCitation's free Cite a source builds the reference from the DOI registry's record, with the version and "[Computer software]" for software; it never writes any part of your report.
References
- Bureau of Labor Statistics. (2026, August 27). Data scientists. Occupational Outlook Handbook, U.S. Department of Labor. https://www.bls.gov/ooh/math/data-scientists.htm
- Government Digital and Data Profession. (2025a, August 29). Data analyst. Government Digital and Data Profession Capability Framework. https://ddat-capability-framework.service.gov.uk/role/data-analyst
- Government Digital and Data Profession. (2025b, August 29). Data scientist. Government Digital and Data Profession Capability Framework. https://ddat-capability-framework.service.gov.uk/role/data-scientist
- Graduate Futures Institute. (2025, May). Data scientist. Prospects. https://www.prospects.ac.uk/job-profiles/data-scientist
- Harvard SEAS. (n.d.). Master's in data science. Harvard John A. Paulson School of Engineering and Applied Sciences. https://seas.harvard.edu/applied-computation/graduate-programs/masters-data-science
- London School of Economics and Political Science. (n.d.). MSc Data Science. https://www.lse.ac.uk/study-at-lse/graduate/msc-data-science
- National Careers Service. (n.d.). Data scientist. https://nationalcareers.service.gov.uk/job-profiles/data-scientist
- Royal Statistical Society. (n.d.). Data Science Professional. https://rss.org.uk/membership/professional-development/data-science-professional/
- Skills England. (2021, June 1). Data analyst (ST0118, Version 1.1). https://skillsengland.education.gov.uk/apprenticeships/st0118-v1-1
- Skills England. (2024, April 24). Data scientist (integrated degree) (ST0585, Version 1.1). https://skillsengland.education.gov.uk/apprenticeships/st0585-v1-1
- University of Bristol. (n.d.). MSc Data Science. https://www.bristol.ac.uk/study/postgraduate/taught/msc-data-science/
- University of Edinburgh. (n.d.). Data Science MSc. https://study.ed.ac.uk/programmes/postgraduate-taught/902-data-science
- Zenodo. (n.d.). FAQ versioning. https://zenodo.org/help/versioning