41297 Chap.1 Business Data, Variables and Measurement Levels
Business Data, Variables and Measurement Levels
Where the module starts
The first two lectures set up everything that follows. Business decisions draw on three kinds of analytics: descriptive work that summarises what happened, predictive work that estimates what comes next, and prescriptive work that recommends an action.
Each depends on a dataset whose rows are observations and whose columns are variables.
What you must be able to say about a variable
This chapter teaches you to describe any column by four answers: its type, quantitative or qualitative; its measurement level, from nominal through ordinal and interval to ratio; its time dimension, cross-sectional or time series; and whether the data are structured.
It also separates a population from a sample and a parameter from a statistic.
Why it pays off later
The level of a variable decides which averages and spread measures are meaningful, so a correct description here prevents the most common error in later chapters: computing a mean for data that can only be ranked.
What this chapter covers
- 01
Descriptive, predictive and prescriptive analytics
- 02
Observations, variables and business statistics
- 03
Population, sample, parameter and statistic
- 04
Nominal, ordinal, interval and ratio levels
- 05
Cross-sectional and time-series, univariate and multivariate data
- 06
Structured and unstructured data
Worked example · free
Classify the columns of a gym membership file
- 1Membership plan: qualitative and ordinal, because Basic, Plus and Premium rank from least to most included.
- 1Home club: qualitative and nominal, a label with no natural order.
- 1Visits in March: quantitative and ratio, a count with a true zero.
- 1Monthly fee: quantitative and ratio, measured in pounds.
- 1Time: cross-sectional, since every row is a different member observed in the same month.
Key terms
- Observation
- One row of a dataset, such as a single customer, store or month.
- Parameter
- A numerical summary of a whole population, written with a Greek letter such as mu.
- Statistic
- The matching numerical summary computed from a sample, such as the sample mean x-bar.
- Ordinal level
- A measurement level whose values rank categories without guaranteeing equal gaps between them.
- Prescriptive analytics
- Analysis that recommends a course of action rather than only describing or predicting.
Business Data, Variables and Measurement Levels FAQ
What is the difference between descriptive, predictive and prescriptive analytics?
Descriptive analytics summarises what has already happened, for example last quarter's sales by region. Predictive analytics uses past data to estimate likely future values. Prescriptive analytics goes further and recommends a decision, such as how much stock to order. The opening lecture introduces all three as the setting for the module's quantitative methods.
How do I tell an ordinal variable from an interval one?
Check whether the gaps between values mean the same amount everywhere. Satisfaction ratings from 1 to 5 rank opinions, but nobody can show the step from 1 to 2 equals the step from 4 to 5, so they are ordinal. Temperature in degrees Celsius has equal steps but no true zero, which makes it interval.
Is a variable stored as numbers always quantitative?
No. Region codes, product category numbers or rating scales are often stored as digits but still record categories or ranks. They are qualitative, nominal or ordinal, and an average of them usually has no business meaning. Decide the type from what the values represent, never from how the spreadsheet stores them.
What is the difference between cross-sectional and time-series data?
Cross-sectional data observe many units at a single point in time, such as 300 stores in one month. Time-series data follow one unit, or a set of units, across successive periods, such as one store's monthly sales over three years. Forecasting methods later in the module need the time-series kind.
Why does the population versus sample distinction matter so early?
Because the symbols and some formulas change with it. A population mean is written with mu and a sample mean with x-bar, and the population and sample variance formulas divide by N and by n minus 1 respectively. Inference, covered later in Semester 1, is the business of moving from a sample statistic to a population parameter.
Assessment move
Open the dataset you are considering for the assessment and build a five-column table: variable name, type, level, unit and time dimension. Fill it before you calculate anything. Then circle every variable at ordinal level or below and note which summaries Chapter 4 allows for it.
Redo the table whenever you add or drop a column, because the seminar form expects the description of the cropped dataset you actually analyse, not the original download.
Working through Business Data, Variables and Measurement Levels in 41297? Sia is AskSia’s AI Statistics tutor — ask any 41297 Business Data, Variables and Measurement Levels question and get a clear, step-by-step explanation grounded in how 41297 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.