Data analytics and insights study guide for the CA Program

Data Analytics and Insights in the CA Program treats data as a business problem first and a technical one second. Eight chapters run from framing a question and preparing the data, through the four analytics types, to visualisation and telling the story. There is no exam. It is assessed by one submission in two parts, a written report with your supporting analysis files worth 80 per cent and a recorded presentation worth 20 per cent. The marks are in the insight and the recommendation, not the analysis.

What you are up against

The technical worry most candidates arrive with is the wrong one. You do not need to code and you do not need a particular tool, and the subject says so plainly. What it asks instead is harder to prepare for by reading: can you look at a business question and work out whether data can answer it, which kind of analysis would, and what someone should do once you have.

That is a judgement subject wearing a technical subject’s clothes. The techniques are taught as concepts rather than as buttons, which means they transfer to whatever your workplace uses, and it also means you cannot revise them by practising in a tool.

Eight chapters, and they run in a sensible order. Frame the problem, prepare the data, then work up through the four analytics types, then make the finding readable and tell it as a story that ends in a decision. The two ends of that sequence are where candidates lose ground: preparation at the start, because it feels like admin, and the recommendation at the end, because it feels like the analysis should speak for itself.

Data analytics and insights is an elective, so you have chosen it. All thirteen CA Program subjects are set out together if you are still deciding the rest.

Where the hours go

Two of the subject’s own signals point in different directions here, and it is worth knowing which one to follow.

The study plan front-loads. It gives framing and data preparation the first two weeks, the descriptive and diagnostic pair the next two, and then compresses prescriptive analytics, visualisation and storytelling into a shared final fortnight. Read as a calendar, that says the early chapters are the heavy ones.

The assessment says the opposite. The marks are in the insight and the recommendation. Two of the six named traps are about stopping too early: stopping at descriptive, and presenting your analysis instead of a recommendation. The chart below follows the assessment, because the assessment is what is graded and the plan is study advice.

Data preparation sitting in the middle needs a word, because the plan gives it two of six teaching weeks. It is genuinely most of the work in a real analysis, and the subject flags skipping it as a trap for good reason: an analysis built on unclean or unsuitable data fails quietly, which is the worst way for something to fail. But it is preparation for the marks rather than the marks themselves, so give it the hours and do not expect it to carry the answer.

The report, and the files that come with it

There is no exam and nothing to pass on its own. One submission in two parts, handed in together in the final week.

The written report is eighty per cent, and it is the only assessment across these electives that asks for more than prose. You submit your supporting analysis files alongside the report, which changes how you work: the analysis has to be legible to someone else, not just correct in your own head. Name things sensibly, keep the steps in order, and leave the working somewhere a reader can follow it.

The report itself is findings, actionable insights and recommendations. Notice the middle word. An insight that is not actionable is an observation, and the subject’s own trap list says presenting analysis instead of a recommendation is where answers go wrong.

The presentation is twenty per cent: a video communicating your insights on the same business problem and how it should be addressed. Same material as your report, said out loud. That is a useful test in itself, because a recommendation that sounds vague when spoken usually was vague on the page.

Choosing the type before touching the data

The four analytics types are a ladder, and each rung answers a different question. Descriptive says what happened. Diagnostic says why. Predictive says what is likely next. Prescriptive says what to do about it.

Learning that ladder is easy. Choosing from it under pressure is the actual skill, and it is the subject’s first named trap: reaching for a technique before framing the business question. The question decides the type, never the other way around.

The rung that catches people is the second one. A relationship in the data is not a cause, and a diagnostic answer that treats a correlation as an explanation has skipped the work rather than done it. Two things moving together is where the investigation starts.

The unglamorous chapter

Data preparation is collection and privacy, cleansing, wrangling, transformation, and storage. It is the part nobody puts in a portfolio and it is most of what a real analysis consists of.

It also carries marks in a way that is easy to miss. Judging whether the data can answer the question is part of the analysis, not a step before it. Data that is incomplete, inconsistently coded, at the wrong grain, or collected for another purpose entirely will still produce output, and the output will look like an answer. That is why the subject calls this failing quietly.

The habit worth building is to write down what you had to do to the data and why, as you go. It costs nothing at the time, it is most of your method section later, and it is exactly the reasoning the analysis files are meant to show.

How to study this subject

Practise the choice, not the technique. Take real business questions, one line each, and name the analytics type each one calls for before touching any data. The subject’s own plan sets this as a task in the week before the assessment, and it is worth doing from the first week instead. It is quick, it needs no tool, and it rehearses the decision every answer starts with.

Work one dataset all the way up the ladder rather than four datasets one rung each. Describing, then diagnosing, then predicting, then prescribing on the same data is how the four types connect, and connecting them is what the written case asks for.

Then practise finishing. Take an analysis you have already done and write the recommendation in three sentences: what you found, what it means, what should happen. Most of the marks live in those three sentences and most candidates spend their preparation time on the part before them.

Traps that cost easy marks

Questions people ask about Data analytics and insights

If this helped, three more will too. Start with the Risk, technology and AI study guide for the subject that treats technology as a risk to be governed, then how to read a question stem for the command verbs that decide what an answer needs, and how to find the topics that carry the marks.