HSC Enterprise Computing Data Science: exam question types and how to answer them

What this guide covers

Data Science is the largest part of Year 12 Enterprise Computing. It asks you to classify data, store and query it, analyse it in spreadsheets and models, and judge how it is collected and used under Australian law. This guide shows how that content becomes exam questions and gives a short method for each type. Try the practice quiz on this page first, then use the sections below to fix anything you missed.

The HSC exam is online and worth 80 marks. NESA's specification lists objective-response items worth about 20 marks (each item 1 to 4 marks) and short-answer items worth about 60 marks. The quiz here uses one-mark multiple choice, which is the simplest objective-response format.

Question types you should expect

1. Classify the data

You are given a variable or a scenario and asked what kind of data it is.

• Quantitative or qualitative? Numbers that label a category (postcodes, 1 to 4 ratings) are qualitative.
• Level of measurement? Work through it in order: is there an order (ordinal or higher)? Are the gaps equal (interval or ratio)? Is there a true zero (ratio)?
• Structured or unstructured? Fixed fields in rows and columns are structured; free text, images and video are unstructured.

See quantitative and qualitative data and levels of measurement and structured, unstructured and alternative data.

2. Pick the right tool or structure

These items describe a need and ask which storage method, database design or spreadsheet feature fits. Match the clue to the tool:

Revise spreadsheet analysis and dashboards, flat-file and relational databases and evaluating data storage methods.

3. Apply the law and ethics to a scenario

A business collects or shares personal information and you must name the obligation or the issue. Keep the jurisdictions straight: the Privacy Act 1988 (Cth), its Australian Privacy Principles and the Notifiable Data Breaches scheme (OAIC) cover most businesses, while NSW public sector agencies are covered by the PPIP Act and HRIP Act (IPC). Indigenous data sovereignty is about who governs data concerning Aboriginal and Torres Strait Islander peoples.

Revise data legislation and Indigenous data sovereignty, ethical, social and legal issues with data and software features affecting privacy and security.

4. Explain how a technology works, then judge it

Big data, data mining, blockchain and machine learning questions usually come in two halves: describe the mechanism, then weigh a benefit against a risk. For blockchain, the mechanism is the chain of hashes agreed by consensus; the limitation is that it cannot prove data was true when it was entered. For machine learning, name supervised or unsupervised learning and say what the model depends on (enough good-quality, representative data).

Revise big data, warehousing and data mining, blockchain for managing and verifying data and machine learning and statistical modelling.

A method for multiple-choice items

1. Read the stem twice and underline the scenario detail that decides the answer (speed of arrival, labelled outcomes, who must be notified).
2. Predict the answer before you look at the options.
3. Eliminate options that describe a real concept that does not fit this scenario. Most distractors are true statements about the wrong idea.
4. For calculation-style items (certainty factors in Intelligent Systems, spreadsheet outputs here), write the working in the scratch space rather than guessing.

Worked example

Question (exam style, 3 marks). A gym app records members' visit times automatically at the door scanner and asks members to rate each class from 1 to 5. Classify each data item and explain one limitation of the rating data.

Answer. Visit times are quantitative and collected passively by computer, because the scanner records them without the member entering anything. The class rating is ordinal qualitative data collected actively, because members choose a ranked category. One limitation is that the gaps between ratings are not equal, so an average rating can mislead; the median or the distribution of ratings is a fairer summary.

Other dot points in this module

Also check data sampling, collection and data quality and curated data and social behaviour. When you are confident with Data Science, move on to the Data Visualisation and enterprise project practice and the Intelligent Systems practice.