Technology & digital skills

Data science that starts with a useful question.

Learn to turn a business question into an analysis someone can inspect, reproduce and use. The aim is to understand the data and the limits of the answer, not simply to produce a model.

This is a learning-path guide, not a confirmed course intake. Ask about current availability, syllabus, delivery format, entry requirements, fees and assessment before applying.

Ask about training availability

Check your starting knowledge.

Basic numeracy, comfort with spreadsheets and a willingness to learn code provide a useful foundation. Ask how a programme introduces statistics, SQL and Python if these are new to you.

Confirm the expected level and any qualification separately. This guide does not advertise a degree or a master's award, even where older promotional material used similar language.

Topics to agree in the syllabus.

A coherent learning path should follow the question through preparation, analysis and evaluation.

Problem framing: a decision, an outcome measure and a useful baseline.
Data understanding: sources, fields, missing values, quality and permitted use.
Exploration: SQL, summaries, visualisation and patterns worth investigating.
Modelling: a suitable baseline before a more complicated method.
Evaluation: appropriate measures, separate evaluation data and error analysis.
Communication: reproducible findings, uncertainty and a recommendation with limits.

A practical project to discuss.

Use a synthetic service dataset to investigate why some tasks finish later than planned. Define what counts as late, inspect missing and inconsistent records, and compare a simple baseline with a candidate prediction method.

Keep information that would only become known after completion out of the prediction inputs. Explain whether the analysis can support an operational decision and what it cannot establish about cause.

Make the analysis reviewable.

The result should be more than a presentation of charts.

A business question and a data dictionary.
A reproducible notebook or script with setup instructions.
A record of quality checks, exclusions and assumptions.
A baseline comparison and an explanation of evaluation choices.
A short decision note describing uncertainty, limitations and the next useful investigation.

Check whether the reasoning transfers.

A learner should be able to explain a surprising result, identify a weak data field and rerun the analysis from the supplied inputs. Ask how changes to the data or business question affect the conclusion.

A high score on one dataset does not prove commercial value or suitability for a different population. Assessment should reward a clear account of what was tested and where the evidence ends.

Start with your
actual requirement.

Tell us the work, people and outcome you have in mind. We’ll confirm current scope and availability before agreeing the next step.