Data, Analytics and AI
Python for data work
Python is where analysis goes when a single query or a workbook can no longer hold the problem: several sources, cleaning repeated every week, a model, a chart that must be rebuilt identically each Monday. The subject here is the working kit rather than the language: pandas and NumPy for tables and arrays, notebooks for exploration, plotting libraries for output, and the habits that stop a notebook turning into an unrepeatable one-off.
Why this topic exists: pandas, NumPy, notebooks and plotting: the working environment of analysis and machine learning (SFIA Data science; the AI pillar's argument for Python).
Python became the working language of data through its ecosystem rather than its design. It was general-purpose, readable and easy to extend with fast libraries written in C, and the scientific community built its tools there. SFIA 9's data science skill speaks of specialised languages and tools without naming any; Python is the one most data teams mean.
The subject here is the kit, not the language. General Python, from variables and functions to packaging, belongs to programming fundamentals. For data work the language matters mostly as the glue between a few libraries and one working environment. Two habits complete the kit in most teams: a separate virtual environment per project, so library versions do not collide, and version control, so a colleague can see what changed between runs.
The kit
- NumPy. Arrays of numbers and fast operations across them. Almost everything else is built on it, which makes its idea of vectorised operations, one instruction applied to a whole column at once, worth understanding early.
- **pandas.** The DataFrame: a table with labelled columns and an index, plus the verbs to filter, join, reshape and group it. Data wrangling is what pandas was made for, and most hours spent in it go on cleaning rather than analysis.
- Plotting. Matplotlib underneath, seaborn and others on top, for the charts that test an assumption during analysis. Presentation-grade data visualisation often moves to a BI tool afterwards.
- Notebooks. Jupyter mixes code, output and notes in a single document, which suits exploration and explaining a result, and suits production badly, because hidden state makes each output depend on the order in which cells ran.
- Modelling libraries. scikit-learn for machine learning and statsmodels for statistical tests, the bridge to the topics that follow.
Notebooks rot
Cells run out of order leave a notebook whose saved output no longer matches its code, and nobody can reproduce last month's figure. Restart and run everything from the top before sharing, pin library versions, treat raw data as read-only, and move any logic that runs weekly out of the notebook into a script or module under version control.
Credentials and the path onward
Two programmes exercise this kit. freeCodeCamp's free Data Analysis with Python certification is built on projects in pandas and NumPy, though it now sits in the archive and is no longer updated. Google's Advanced Data Analytics certificate uses Jupyter notebooks throughout its statistics, regression and machine learning modules. Neither checks the reproducibility habits above, and those are what a reviewer notices first. The beginner's other mistake is looping over rows one at a time in pandas: it works on a sample and crawls on a real table, and learning the vectorised equivalent is the first genuine speed-up. From here the path runs to statistics and experimentation and machine learning, and, once datasets outgrow one machine, to data engineering. The Python tool page covers the language itself.
Next to this topic
- Data analysis fundamentalsAsk, prepare, process, analyse, share: the analyst's workflow as Google's programme and DP-900's core data concepts teach it, before any tool (SFIA Data analytics).
- SQL for analysisQuerying, joining, aggregating and windowing data is the analyst's first real tool and the language of every warehouse (DP-900 relational concepts; DMBOK Data Storage and Operations; the programming field owns database design).
- Data visualisation and business intelligenceDashboards, reports and the semantic models behind them: PL-300's four skill areas and DMBOK's Data Warehousing and BI area (Tableau, Power BI, Looker).
- Statistics and experimentationProbability, inference, regression and A/B tests separate a chart from a claim (Google's Advanced Data Analytics; SFIA Data science).
- Data engineering and pipelinesIngestion, transformation, warehouses, lakehouses and orchestration (DMBOK Data Integration and Interoperability; DEA-C01; DP-700; Databricks; Spark, Airflow, dbt).
- Data governance, quality and privacyOwnership, definitions, metadata, master data, quality rules and lawful use are DMBOK's centre and half its wheel; nobody hires for them until the numbers disagree.
- Machine learningSupervised and unsupervised learning, model evaluation, deep learning and the discipline of not fooling yourself (SFIA Machine learning; AIF-C01; Azure AI Fundamentals).
- Generative AI, LLMs and agentsFoundation models, prompting, retrieval, fine-tuning and agents: the newest layer of the field and the one AIF-C01 and Azure AI Fundamentals now test (the AI portal reads the concepts in order).
- Natural language processing and computer visionText and images are the two applied fields where classic machine learning met deep learning first, and where most production AI still lives (SFIA Machine learning; Azure AI Fundamentals AI-901 workloads).
- Responsible AI and AI governanceBias, transparency, privacy, safety and the frameworks that make them work (NIST AI RMF; AIF-C01's responsible-AI domain; SFIA 9's AI and data ethics skill; the EU AI Act).
- MLOps and AI engineeringGetting a model or an LLM application into production and keeping it there: deployment, monitoring, evaluation, cost (SFIA Machine learning and Data engineering; Applied Skills in Foundry; MLOps in the AI portal).
Concepts to know
Glossary entries with the reason each one matters here.
- Python
The topic's language.
- Data Wrangling
pandas exists for this.
- Data Visualization
matplotlib and its successors.
Certifications that test it
Vendor exams and free certificates; facts, cost and the preparation path are on each page, and the certifications hub has them all.
Tools of the trade
- Python
pandas, NumPy and notebooks run on it.
- pandas and NumPy
The tabular data library of Python.
Frequently asked
- Python or R for data work?
- For most analyst and data science roles, Python, because the same language carries through to machine learning and production code. R remains strong in statistics and academic research; the statistics topic explains where it still leads.
- How much Python do I need before starting pandas?
- Less than people expect: variables, lists and dictionaries, functions, loops and the patience to read an error message. pandas teaches its own idioms, and many of them replace the loops a beginner would otherwise write.
- Is spreadsheet experience wasted once I learn pandas?
- No. Plenty of colleagues will still want results in a workbook, and spreadsheet thinking, with its columns, filters and pivot tables, maps directly onto DataFrame operations. Spreadsheets remain the quickest tool for a small one-off check.
Courses in the directory
205 courses are filed here; the top 6 by our ranking, details and the provider link on each course page.
Last reviewed 26 September 2026 · Getting Digital