22576 Chap.1 Business Analytics, Data Types and Data Structures
Business Analytics, Data Types and Data Structures
The first week sets the vocabulary the rest of the subject relies on. Business analytics is framed as turning data into insight for decisions, in three kinds: descriptive work on what happened, predictive work on what will happen, and prescriptive work on what to do.
The chapter then sorts variables by measurement scale and by how software stores them, and distinguishes cross-sectional, time series and panel data, including why a stacked panel suits modelling better than a wide one.
These ideas are tested early.
Quiz 1 in the Week 4 tutorial covers Weeks 1 to 3, and its conceptual questions often turn on a single classification, such as whether a coded rating is ordinal or whether a table is a panel. The same vocabulary returns in the group report, where every variable you clean or chart needs the right type before it can be summarised, merged or modelled.
What this chapter covers
- 01
Business analytics as a decision-focused process
- 02
Descriptive, predictive and prescriptive analytics
- 03
Nominal, ordinal, interval and ratio scales
- 04
Categorical, numerical, date and text types in software
- 05
Structured versus unstructured data
- 06
Cross-sectional, time series and panel layouts
- 07
Stacked (long) versus unstacked (wide) panels
Worked example · free
Classify four columns from a café's sales file
- 1Payment method is categorical nominal: the categories are labels with no order.
- 1Customer rating is categorical ordinal: the stars are ordered, but the gaps between them are not guaranteed equal.
- 1Number of items is numerical discrete: it is a whole-number count.
- 1Order total is numerical continuous; in Power Query check it is stored as a decimal number and not as text.
Key terms
- Prescriptive analytics
- Analysis that recommends what to do, building on descriptive and predictive results.
- Ordinal variable
- A categorical variable whose categories have a natural order, such as a satisfaction level.
- Ratio scale
- A numerical scale with equal spacing and a true zero, such as sales or age.
- Panel data
- Data that follow many entities across several time periods, combining cross-sectional and time series data.
- Unstructured data
- Data without a predefined format, such as free text or images, that cannot be stored directly in rows and columns.
Business Analytics, Data Types and Data Structures FAQ
What is the difference between descriptive and predictive analytics?
Descriptive analytics summarises past performance and looks for root causes, answering what happened and why. Predictive analytics forecasts what will happen. In this subject the descriptive weeks come first, covering wrangling, summary statistics, clustering and tests, and the predictive weeks follow with regression and classification models built in KNIME.
Why is a phone number or postcode treated as text?
Because it identifies something rather than measuring it. Nobody adds or averages postcodes, and numeric storage can strip leading zeros. Software often guesses these columns as numbers, so the tutorial practice of checking and reassigning each column's type before analysis protects both merges and summaries.
Why is the stacked layout preferred for modelling?
A stacked, or long, panel keeps each variable in a single column, with time as its own variable. A wide layout spreads one variable across several columns, one per period, which breaks the one-variable-per-column rule and stops a model from treating that variable as a single input.
Why does Power Query sometimes assign the wrong data type?
It guesses each column's type from the values it sees, so identifiers made of digits become numbers and flags written as single letters stay as text. The Week 2 tutorial has you inspect the type icon on every column and reassign the wrong ones before any cleaning or analysis begins.
What makes data structured rather than unstructured?
Structured data follow a fixed format, usually rows and columns, so software can sort, filter and summarise them directly. Unstructured data, such as reviews, transcripts or images, have no predefined format; they are harder to analyse and usually need converting into structured fields first, a task where AI tools can help.
Assessment move
Make a two-column list of every variable in a tutorial dataset with its scale and its software type, and note where they disagree. Practise reshaping one small panel between wide and long by hand, counting rows and columns before and after, until the arithmetic is automatic.
Finally, test yourself on the three kinds of analytics by labelling each technique in the subject program as descriptive or predictive, and note where recommendations would become prescriptive.
Working through Business Analytics, Data Types and Data Structures in 22576? Sia is AskSia’s AI Business Analytics tutor — ask any 22576 Business Analytics, Data Types and Data Structures question and get a clear, step-by-step explanation grounded in how 22576 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
