Instructor Notes
About This Course
Methods for Data Science is the data-methods half of the foundation pathway. The teaching goal is not to turn learners into statisticians or programmers. It is to improve judgement about data collection, organisation, interpretation, and communication.
The course works best when learners leave with:
- a clearer idea of what makes data fit for purpose
- stronger habits for inspecting and organising tabular data
- more confidence critiquing visualisations and summaries
- a practical sense of how findings become decisions, actions, and feedback
Recommended Teaching Approach
- Keep examples concrete and low-risk.
- Ask learners to connect each idea to a real dataset, chart, or report.
- Treat data quality as design work, not just cleaning work.
- Separate what the data shows from what the learner thinks should happen next.
- Use debriefs to connect analysis choices to real-world consequences.
Before Teaching
- Review the Course Overview, Learner Setup, and the four linked episodes. Confirm the session plan against the episode agendas; optional exercises below do not alter those timings.
- Check the original and prepared student spreadsheets open, and test the browser activity in Making Sense of Data on the teaching network.
- Keep the fictional request worksheet available as an access fallback. It includes source notes and worked checks, so a pair can participate on paper.
- Prepare a shared-screen or paired-access route for learners without spreadsheet access. No coding installation is required.
- Use public or fictional examples. Invite a description of a real research question rather than requesting confidential records.
Facilitation and Checks
For a five-minute optional task, allow one minute to inspect, two minutes to attempt, one minute to check, and one minute for two pairs to report. Count the debrief inside the five minutes. Use the Discussion questions to connect a task to learners’ own research.
| Watch for | Ask or demonstrate |
|---|---|
| Removing rows merely because values match | “What does a row represent, and what source confirms this is a duplicate?” |
| Replacing an impossible value with a reasonable guess | Flag the value, check the source, and log a correction only when supported |
| Treating blanks as zero | State how many values contribute to the calculation and what is unknown |
| Removing a large valid value | Compare its source and context; show its effect on mean, median, and spread |
| Treating an association as a cause | Ask for competing explanations and what further design or evidence is needed |
| Calling a recommendation an impact | Name an owner, a feedback measure, and a check for unintended effects |
Ask learners to keep the original data, working copy, and change log. An apparently cleaner table should still be traceable to its source.
Session-Level Emphasis
Collecting the right data
Use Collecting the Right Data to show that analysis quality starts before analysis begins. Emphasise question design, source quality, variables, formats, and fairness.
Organising and exploring data
Use Organising and Exploring Data to make messy data visible. Learners should practise spotting inconsistencies, standardising structures, and asking early exploratory questions.
Making sense of data
Use Making Sense of Data to build visual and interpretive judgement. Emphasise that charts are arguments, not neutral decorations.
Insight to impact
Use Insight to Impact to connect findings to audiences, action, feedback, and iteration. Keep asking what would change because of the insight.
Consolidation and Follow-On Learning
The learner Additional Material has four short tasks matched to these lessons, including worked answers. The Reference provides reusable checks. Use Instructor Additional Material and Extra Exercises for optional extensions matched to the course outcomes.
For learners ready to implement the same checks in code, signpost Applied Python or Applied R. For foundational storage or governance questions, signpost Essential Digital Skills.