How data science has changed
Data science in 2026 looks significantly different from the peak hype of 2018-2022. The role has bifurcated into two distinct tracks: applied data science (building and deploying models that solve business problems in production) has merged increasingly with ML engineering, requiring stronger software engineering skills than early data science roles. Analytical data science (using data to understand business performance, find opportunities, and drive decisions) has similarly merged with business intelligence and data analytics, requiring sharper business acumen and communication skills alongside technical capability.
The "unicorn" data scientist of the early AI boom (expert statistician, expert programmer, expert communicator, and domain expert all in one) never really existed at scale, and the role has matured into more specialised tracks. AI tools (including generative AI) have automated some of the routine coding and data manipulation tasks that junior data scientists spent significant time on, compressing the time between question and insight but also raising the bar for what constitutes valuable data science work.
Skills that matter in data science in 2026
The foundation remains: Python (and/or R for statistical work), SQL, statistics and probability, machine learning fundamentals, and the ability to communicate findings clearly. What has become more important: ML engineering skills (deploying models, building pipelines, MLOps); working with large language models and the ecosystem of tools around them (LangChain, vector databases, evaluation frameworks); causal inference and experimental design (A/B testing methodology, understanding causation vs. correlation rigorously); and strong business communication that translates analytical findings into decisions. What has become less differentiating: knowing how to use sklearn for standard classification and regression tasks, because this is now commodity knowledge that AI tools assist with substantially.
Is data science still worth pursuing?
Yes, with clear eyes about the direction of the field. Data science is a strong career with good compensation and continued demand, but the most valuable data scientists in 2026 are those with genuine mathematical and statistical depth, strong ML engineering capability, and the communication skills to connect technical work to business outcomes. The "entry-level data scientist who primarily works in notebooks doing ad hoc analysis" position is under more pressure than it was five years ago. The senior, specialised, and ML engineering end of the data science career ladder is strong. UK salaries range from £40,000-£55,000 at junior level to £80,000-£120,000 at senior level, with ML engineering specialisation often commanding a premium.