Python for Data Analysts: What UK Employers Actually Expect in 2026

A data analyst uses Python to clean, reshape and check data that does not already sit in a tidy database: exported files, repeated cleaning steps, and questions a dashboard cannot answer cleanly. In the UK, as of October 2026, it is a common skill on job ads, and it is not a substitute for SQL.

Courses often sell Python as the skill that makes you a data person. UK hiring does not work like that. This guide uses one public source, says what the word "Python" means on an analyst advert, and sets an order of study that matches how people actually get a first role. If you are still mapping the wider path, start with how to become a data analyst and our essential skills breakdown.

The short answer

• SQL comes first. On UK permanent ads with "Data Analyst" in the title, SQL is named more often than Python (42% versus 30%).
• "Python" means tables, not websites. Employers want pandas-style cleaning and grouping. They are not asking for Django, Flask, or a machine learning portfolio.
• Power BI is named more often than Python (36% versus 30%). A BI tool belongs in the plan before a deep dive into code.
• R is a niche on these ads (2.4%). Python is the second language to learn unless a specific team already uses R.
• Six to ten weeks, part time, is a realistic add-on once your SQL is already solid. It is not a year of computer science.

What a UK advert means by Python

IT Jobs Watch tracks skills that appear alongside a job title in UK permanent vacancies. For the six months to 11 October 2026 it lists 1,125 ads with Data Analyst in the title. The median salary on the ads that quote one is £54,000. The skill counts below are co-occurrences: an ad can name several tools, so the shares do not add up to 100%.

Read 30% carefully. It is not "most analyst jobs require Python", and it is not "Python is optional, ignore it". About three in ten of these ads name it. The ads that do not name it can still hand you a messy export on a Tuesday. The ads that do name it rarely spell out a library. Pandas shows up as its own line on a handful of postings, because recruiters write "Python" and assume you know they mean tabular data.

Two limits on this source, so you do not over-read it. First, it only covers ads whose title contains Data Analyst. Insight analyst, reporting analyst and MI analyst sit outside the count. Second, it does not split salary by skill, so this article will not invent a Python pay premium. For the pay bands themselves, see Data Analyst salary in the UK.

The broader developer picture is different, and worth one sentence of context. In the 2025 Stack Overflow Developer Survey, 57.9% of respondents had used Python, up 7 percentage points on 2024. That is professional developers in general, not UK analyst ads. It explains why libraries, hiring managers and AI coding tools have settled on Python. It does not change the order in the table above.

Excel, SQL or Python

You will use all three in a real analyst job. The mistake is treating them as a ranking of prestige. They answer different shapes of question.

SQL is the language of the database the company already trusts. Our SQL guide for data analysts covers the level UK interviews actually test. Power BI versus Tableau is a separate choice, driven by which tool the employer already pays for: the Power BI vs Tableau comparison uses the same kind of job-board evidence.

Python earns its place when the data is not query-ready. A CSV from a vendor with three date formats. A monthly export you would rather not clean by hand again. A join between a warehouse extract and a spreadsheet that will never be loaded into the warehouse. A check that the row count after a filter is the row count you expected. That is analyst Python. It is closer to a careful spreadsheet than to software engineering.

Python or R

For a UK data analyst job search in 2026, learn Python. R appears on 27 of the 1,125 ads in the IT Jobs Watch window, Python on 338. That gap is wide enough to settle the default.

R still has a real home in some statistics-heavy teams: parts of pharma, some academic groups, some government analysis units. If the advert names R, learn enough R to do that job. Do not learn both languages before you have a first offer. Two half-learned tools lose to one tool you can use on a messy file while someone watches.

Libraries to learn, and libraries to skip

Job ads say "Python". The work is a short list of libraries. Learn them in this order, and stop when the list runs out.

1. pandas

This is the library. Reading a file, checking how many rows and columns you actually have, finding nulls, filtering, grouping, merging two tables, writing the result back out. The pandas getting started guide is enough documentation. You do not need a second course that re-teaches the same ten methods with new variable names.

The shape of a useful session looks like this, not like a framework tutorial:

import pandas as pd

orders = pd.read_csv("orders.csv")
print(orders.shape)
print(orders.isna().sum())

monthly = orders.groupby("month", as_index=False)["revenue"].sum()

If shape or the null count surprises you, stop and look. That pause is the skill. The groupby is the easy part.

2. One charting library

matplotlib or seaborn. Pick one. You are drawing a chart to check a pattern while you explore, not to publish a dashboard. The dashboard still belongs in Power BI or Tableau, where your colleagues can filter it without opening your notebook.

3. Enough core Python to not get stuck

Variables, lists, dictionaries, a function, a loop, and reading an error message. That is the language underneath pandas. If the syntax itself is the part that feels blocked, the free Python Foundations crash course is seven short daily lessons on exactly that, before you touch a dataset. It is a start. It is not a portfolio.

Leave these until the role changes

• scikit-learn, TensorFlow, model training. That is a data scientist or AI path. Our data scientist roadmap is the right article if that is the job you actually want, and the analyst, scientist and engineer comparison is the right article if you are not sure. If the work you want is pipelines rather than charts, read Python for data engineers.
• Django, Flask, FastAPI. Web applications. Analyst interviews do not ask for them.
• LeetCode-style algorithms. Some software roles test this. Analyst take-homes almost never do. They hand you a dataset and a question.

What job-ready Python looks like

For a first UK Data Analyst role, job-ready means you can do the following on a file you have not seen before, without a tutorial open beside you:

• Load a CSV or Excel file and say how many rows survived the import.
• Find missing values, duplicated keys, and columns that arrived as text when they should be numbers or dates.
• Filter, group, and merge two tables, then explain what the merge did to the row count.
• Write the cleaning steps as a function you could run again next month.
• State one finding in a sentence a manager could use, and name the assumption underneath it.

You do not need classes, decorators, or a packaged library. You do need to notice when a merge duplicated rows. That single check separates people who have followed a pandas playlist from people who can be left with a file.

How long it takes

Assume you already write intermediate SQL: joins, aggregates, a window function, and a healthy suspicion of nulls. From there, six to ten weeks at five to eight hours a week is enough to reach the list above, if those hours are spent on real tables and not on watching someone else type.

A workable split is two weeks on core syntax if you need it, four weeks on pandas against messy public datasets, and two weeks turning one of those datasets into a short project you can explain out loud. If you are starting from zero, do not begin here. Do SQL first, then come back. People who reverse that order can produce a chart and still fail the technical screen.

AI is now named on 261 of the same 1,125 ads (23.2%). Use a coding assistant to draft a pandas step, then check the row count yourself before you trust the number. The failure mode, and what UK employers still expect you to do unaided, is covered in using ChatGPT and Copilot as a data analyst.

Mistakes that waste months

Learning Python before SQL

SQL is on more ads, and it is what the interview screen usually opens with. Python on top of shaky SQL reads as a course completion, not as an analyst.

Studying syntax and never opening a dataframe

Loops and classes feel like progress because they are marked. Analyst work is a table with a broken date column. If your practice data is already clean, you are practising the wrong thing.

Using a competition notebook as a portfolio

A leaderboard score does not tell a hiring manager whether you can explain a revenue drop. One project that starts from a dirty file, states its assumptions, and ends in a decision is worth more. The README rules that turn that project into interviews are in how to build a data analyst portfolio.

Trusting generated code you cannot check

A model will happily group the wrong column and hand you a tidy chart. If you cannot predict the row count before you run the cell, you are not ready to use the assistant on someone else's numbers.

Showing Python in a portfolio

One project is enough. Give it these pieces, in this order, in the README:

1. The question, in one sentence, and who would act on the answer.
2. Where the file came from, and what was wrong with it.
3. The cleaning steps, including anything you chose to drop, and why.
4. The finding, with the table or chart that supports it.
5. What you would not conclude, because the data cannot carry it.

Put the notebook or script in a public GitHub repository. Do not hide the messy intermediate steps. The messy steps are the evidence that you did the analyst's job rather than the tutorial's job.

Frequently asked questions

Read next

• SQL for Data Analysts: The Honest Beginner's Guide
• Power BI vs Tableau: Which Should You Learn First in 2026?
• How to Build a Data Analyst Portfolio That Actually Gets UK Interviews