Data Scientist

• ·Work out what question is actually being asked, which is frequently not the question that was asked.
• ·Find, clean and reconcile data — consistently reported as the majority of the job, and the part nobody advertises.
• ·Build statistical and machine learning models to predict or explain something.
• ·Validate whether a model actually works on data it has not seen, and resist the temptation to believe it too early.
• ·Design and analyse experiments, so that a change can be credited or blamed honestly.
• ·Visualise and communicate findings to people who will make decisions on them.
• ·Deploy models into production and monitor whether they degrade — increasingly part of the role rather than someone else's job.
• ·Say when the data cannot answer the question, which is the most valuable and least popular thing you will do.

If any of those systems is unfamiliar — or you have not chosen between them yet — the exams and qualifications section covers what each one is, which subject inside it opens which degree, and when to sit what.

This deserves a blunter answer than it usually gets. Data science was heavily promoted for a decade, and the supply of people calling themselves data scientists grew faster than the number of genuine positions. Entry-level competition is now substantial, and many advertised roles are really analytics or reporting jobs with a fashionable title. Experienced people who can both model and engineer remain in demand. The field is real; the hype around how easy it is to enter was not.

Provenance for this sectionestimatedestimatedInferred by reasoning, not measured. The basis is given below.Reasoned from published national data and AI policy, the absence of licensing requirements, and the structural scarcity of experienced practitioners. No occupational employment or wage statistic for Saudi data roles was obtainable.

1. Analyst or junior data scientistyears 0–2reportedreportedConsistently reported across multiple independent credible accounts. Not a measured statistic.Based on 4 independent accounts.Early roles consistently described as dominated by data preparation, reporting and supervised analysis.
2. Data scientistyears 2–5reportedreportedConsistently reported across multiple independent credible accounts. Not a measured statistic.Based on 4 independent accounts.Independent ownership of analyses and models consistently described as arriving after a first couple of years.
3. Senior data scientistyears 5–10reportedreportedConsistently reported across multiple independent credible accounts. Not a measured statistic.Based on 3 independent accounts.Senior level consistently described as marked by choosing which problems are worth solving and mentoring others, rather than by technical depth alone.
4. Principal, lead, or head of datayears 9–18estimatedestimatedInferred by reasoning, not measured. The basis is given below.Inferred from the standard divergence between deep technical and management tracks in data organisations, which occurs after senior level rather than at a defined point.

Product analytics and experimentation

Designing and interpreting experiments. Closest to the scientific method as actually practised in industry.

Machine learning

Predictive modelling in production. Overlaps heavily with engineering.

Causal inference

Working out what actually caused what. The most intellectually demanding and least common specialism.

Business intelligence and analytics

Reporting and decision support. Easier to enter, lower ceiling, and much of the market.

Research science

Advancing methods. Usually requires a doctorate and sits in a small number of employers.

Demand: growingestimatedestimatedInferred by reasoning, not measured. The basis is given below.Inferred from the occupation's recorded employment and its recent emergence as a category, combined with consistently reported saturation at the entry and reporting end. No published occupational projection was obtainable.

262,440 data scientists recorded in the US, in an occupation that barely existed as a distinct category twenty years ago. Organisations continue to accumulate data faster than they can use it. The important qualification is that headline growth conceals a split: demand for people who can genuinely model and engineer is strong, while the reporting-adjacent end is crowded and increasingly automated.

If you like this, consider

• Data EngineerBuilding the pipelines analysis depends on. Less glamorous, consistently easier to get hired into.
• Machine Learning EngineerTaking models into production, with more engineering and less analysis.
• StatisticianThe same foundations with more rigour and less tooling churn.
• ActuaryComparable mathematics with a protected credential and far more predictable progression.
• Quantitative AnalystThe financial-markets end, with higher pay and worse hours.

• See the rest of this fieldWhat else sits near this one.
• Compare it against somethingSide by side on the things that differ.
• Careers you've never heard ofThe point of this site is that the list is longer than you think.
• Back to the whole mapThirty fields, and most of them are unfamiliar.