Choose a unit
Units are listed in teaching order for the trimester. Weeks 3 and 4 each carry two sessions, and Week 7 is the mid-trimester break. The unit code in each page address (unit-05, unit-08) is the module's own numbering and does not follow the week order. Each unit takes roughly 45 to 90 minutes end to end, and every unit has a NotebookLM study companion you can question afterwards.
The Agentic Analytics Workplace
The six-stage pipeline, what agents get wrong, silent degradation, and graduated autonomy.
Directing and Supervising AI Agents
Four prompt patterns, five governance questions, the platform spectrum, and G.R.O.W.T.H.
Data Structure and Tidy Thinking
Acquisition and provenance, the tidy-data rule, data types, the pitfalls that bite, and your project launch.
Missing values, duplicates, category mappings, joins, the verification note, and the Kopi Kembali join lab.
Mean against median, spread, skew, segmentation, and the pivot-table check.
Correlation and causation, confounders, sampling bias, real or noise, A/B boundaries, and the three-question audit.
Web, Social and Campaign Analytics
The four metric families, funnel leaks, CPA and ROAS, traffic quality and attribution.
Web Metrics and Google Analytics II
What the tool counts as a user and a session, dimensions and metrics, event architecture, GAIQ preparation.
Data Studio I: the perception hierarchy, five distortions, proportional ink, accessibility and uncertainty.
Data Studio II: one question, three charts, one filter, visual hierarchy, and your portfolio dashboard.
Audience first, one message, the Pyramid Principle, three structures, framing against distortion.
Chained pipelines, checkpoints, reading generated code before it runs, governance and hard controls. The supervised quiz sits in this week’s session.
Time-Series Analysis and Forecasting
Trend and seasonality, seasonally adjusted series, the baseline rule, honest holdouts and break conditions.
From Insight to Campaign Recommendation
The claim ladder, fairness trade-offs, the premortem, overriding the agent, the rubric, and the group pitch.
Assessment at a glance
Individual Analytics Portfolio 30% (assembled across the trimester; unmarked formative checkpoint in Week 6) · Google Analytics certification 5% (certificate due Week 8) · Quiz 25% (supervised, in the Week 11 session) · Group campaign project 40% (presented in the Week 13 seminar, with AI-use disclosure and peer evaluation). Submissions run through xSiTe; your team’s living artefacts live in Google Drive.
How to use these
Each unit follows the same shape: read the journey, use the toolbox on your own data, build the deliverable, then take the knowledge check. The Build section of each unit produces the artefact that unit contributes to your Individual Analytics Portfolio or your team’s orchestration log, and you can copy or export it.
Ask the notebook opens that unit's study companion in NotebookLM: a chat grounded in the unit's slides, facilitator guide, lab worksheet, knowledge check, this app, the refresher deck, the research the unit cites, and the module syllabus, project brief and rubrics. Ask it to explain something again, to test you, or to check your reasoning. It will not draft your assessed work, and it answers with citations so you can follow anything back to the source. Each one also holds a briefing document and a study guide.
Progress, checklists and quiz answers are stored in your browser and, once you sign in, synced to your BestCourses.sg learner account so you can resume on any device. Signing in also records your course activity (chapters completed, quiz and exam results, and time spent) which teaching staff use to see cohort progress and support anyone falling behind. Usage analytics run through Google Analytics and PostHog. The text you type into builder fields is not collected, only that you completed them. Full details are in the BestCourses.sg privacy policy.
NotebookLM is a Google service and sits outside all of this: what you type into a notebook goes to Google, so treat it as you would any other cloud tool and keep personal data out of it.