I85PREP
⌘ K
EXIN Data Analytics Foundation study materials
EXIN Data Analytics Foundation study resources

EXIN Data Analytics Foundation practice questions

One question at a time. Build your EXIN Data Analytics Foundation understanding.

First 50 of 200 questions. Each one lists the choices, the correct answer, and an explanation. The checker at the top of this page saves progress in this browser when it is available.

Continue in the practice tool above to load more questions, 50 at a time.

Question 1Data Analytics Fundamentals

In data analytics, what does the term 'structured data' most precisely describe?

  1. Raw text collected from social media posts and reviews
  2. Data organised into a predefined model such as rows and columns with typed fields Correct answer.
  3. Any data stored in a computer file regardless of its internal organisation
  4. Data that has already been visualised in a dashboard

Explanation

Structured data conforms to a predefined schema, typically tables with named, typed columns, which makes it directly queryable with SQL. Unstructured and semi-structured data lack such a rigid model and need extra processing before analysis.

Question 2Data Analytics Fundamentals

A retailer records each sale with a timestamp, product code, quantity and price. Which analytics type answers the question 'how many units did we sell last month?'

  1. Descriptive analytics Correct answer.
  2. Prescriptive analytics
  3. Cognitive analytics
  4. Predictive analytics

Explanation

Descriptive analytics summarises what has already happened, such as totals, averages and trends over a past period. Predicting future demand or recommending an action goes beyond summarising historical facts.

Question 3Data Analytics Fundamentals

Which statement best captures the relationship between data, information and knowledge?

  1. Data, information and knowledge are three interchangeable words for the same concept
  2. Knowledge is the raw input and data is the output of analysis
  3. Information is raw symbols and data is the meaning assigned to them
  4. Data are raw facts, information is data processed into context, and knowledge is understanding applied to information Correct answer.

Explanation

The DIK hierarchy treats data as unprocessed facts, information as data given context and meaning, and knowledge as the understanding that lets people act on information. Each level adds interpretation to the one below.

Question 4Data Analytics Fundamentals

What is the main purpose of a data warehouse in an analytics environment?

  1. To replace the need for any data governance policy
  2. To run the daily transaction processing of an operational system
  3. To store integrated, subject-oriented historical data for analysis and reporting Correct answer.
  4. To hold only unstructured multimedia files

Explanation

A data warehouse consolidates data from multiple source systems into a subject-oriented, integrated, time-variant store optimised for querying and reporting. Operational transaction processing remains the job of source systems such as ERP or CRM databases.

Question 5Data Analytics Fundamentals

Which pair correctly matches a data scale with an operation that is valid on it?

  1. Interval scale: computing a difference between two values Correct answer.
  2. Ordinal scale: computing the ratio between two ranks
  3. Nominal scale: computing the average of assigned category codes
  4. Ratio scale: ordering values without any defined zero

Explanation

Interval scales have equal distances between values but an arbitrary zero, so differences such as temperature change are meaningful while ratios are not. Nominal data only supports equality, ordinal data supports ordering, and ratio data has a true zero.

Question 6Data Analytics Fundamentals

A dataset of customer ages contains the values 18, 22, 25, 29 and 91. Which measure of central tendency is most affected by the value 91?

  1. The mean Correct answer.
  2. The interquartile range
  3. The median
  4. The mode

Explanation

The mean uses every value in its calculation, so a single extreme observation pulls it upward. The median and mode are positional or frequency-based and stay comparatively stable in the presence of outliers.

Question 7Data Analytics Fundamentals

A manager wants a single number describing how spread out delivery times are around their mean. Which statistic fits that need?

  1. Correlation coefficient
  2. Median
  3. Standard deviation Correct answer.
  4. Mode

Explanation

Standard deviation expresses dispersion in the same units as the data and is derived from the average squared deviation from the mean. The median and mode are central-tendency measures, and correlation measures association between two variables.

Question 8Data Analytics Fundamentals

A dataset contains missing values in the 'income' column. Which action represents imputation?

  1. Replacing each missing income with an estimated value such as the column median Correct answer.
  2. Leaving the missing cells empty and excluding the column from all reports
  3. Changing the column data type from numeric to text
  4. Deleting every row that has a missing income

Explanation

Imputation fills missing entries with estimated values derived from the observed data, such as the mean, median or a model-based prediction. Deleting rows is listwise deletion, which reduces sample size instead of filling gaps.

Question 9Data Analytics Fundamentals

An analyst joins a sales table to a customer table on customer ID and finds that some sales rows disappear from the result. Which join type would retain every sales row even when no matching customer exists?

  1. Right outer join with customers as the right table
  2. Left outer join with sales as the left table Correct answer.
  3. Cross join
  4. Inner join

Explanation

A left outer join keeps all rows from the left table and fills unmatched right-side columns with nulls, so no sales record is lost. An inner join keeps only matching keys, which explains the disappearing rows.

Question 10Data Analytics Fundamentals

What distinguishes a primary key from a foreign key in a relational database?

  1. A primary key may contain duplicate values, while a foreign key must be unique
  2. A foreign key must always be numeric, while a primary key may be any type
  3. A primary key and a foreign key are two names for the same constraint
  4. A primary key uniquely identifies each row in its own table, while a foreign key references a key value in another table Correct answer.

Explanation

A primary key enforces uniqueness and non-nullness for row identification within a table. A foreign key stores a reference to a key in another table and enforces referential integrity between the two relations.

Question 11Data Analytics Fundamentals

Which activity belongs to the 'prepare' phase of the OSEMN data science process rather than to 'explore'?

  1. Plotting a histogram of a numeric variable
  2. Parsing dates and standardising text categories across sources Correct answer.
  3. Testing a hypothesis about group differences
  4. Reporting final findings to stakeholders

Explanation

In OSEMN (Obtain, Scrub, Explore, Model, Interpret), scrubbing or preparing data covers cleaning, parsing and standardising values. Plotting and hypothesis testing belong to exploration, while reporting belongs to interpretation.

Question 12Data Analytics Fundamentals

A company wants to group customers into segments based on purchasing behaviour without any predefined segment labels. Which technique is appropriate?

  1. Linear regression
  2. A/B testing
  3. Hypothesis testing on a single mean
  4. Cluster analysis Correct answer.

Explanation

Clustering is an unsupervised learning technique that discovers natural groupings in data when no target label exists. Regression and hypothesis testing require a defined outcome variable, and A/B testing compares predefined variants.

Question 13Data Analytics Fundamentals

In a supervised classification model, what is the role of the training set?

  1. It provides labelled examples from which the model learns the mapping to the target class Correct answer.
  2. It stores the final production predictions for auditing
  3. It holds data deliberately excluded from learning to estimate generalisation
  4. It defines the database schema used by the model

Explanation

The training set contains labelled examples that the algorithm uses to fit its parameters. A separate test or validation set is held back to estimate how well the learned model generalises to unseen data.

Question 14Data Analytics Fundamentals

A model achieves 99% accuracy on a fraud dataset where only 1% of transactions are fraudulent. Why is accuracy a misleading metric here?

  1. Because accuracy can only be computed for regression models
  2. Because accuracy ignores the number of records in the dataset
  3. Because a trivial model predicting 'not fraud' for every case would also score about 99% Correct answer.
  4. Because fraud data can never be classified

Explanation

With severe class imbalance, a naive classifier that always predicts the majority class reaches high accuracy while detecting no fraud at all. Metrics such as precision, recall or the F1 score reveal performance on the minority class.

Question 15Data Analytics Fundamentals

Which statement describes a Type I error in hypothesis testing?

  1. Failing to reject a true null hypothesis
  2. Choosing the wrong sample size for the test
  3. Accepting a false alternative hypothesis as true
  4. Rejecting a true null hypothesis Correct answer.

Explanation

A Type I error occurs when the null hypothesis is true but the test rejects it, a false positive controlled by the significance level alpha. A Type II error is the opposite: failing to reject a false null hypothesis.

Question 16Data Analytics Fundamentals

A correlation coefficient of -0.85 between advertising spend and website visits in a given period indicates what?

  1. The data contain an arithmetic error because correlations cannot be negative
  2. There is no relationship between the two variables
  3. Advertising spend causes visits to fall
  4. There is a strong negative linear association between the two variables Correct answer.

Explanation

The coefficient ranges from -1 to +1, and -0.85 signals a strong negative linear association: higher values of one variable tend to accompany lower values of the other. Correlation quantifies association and cannot by itself establish causation.

Question 17Data Analytics Fundamentals

What is the principal purpose of a data governance framework?

  1. To replace data analysts with automated tools
  2. To decide which visualisation tool the marketing team buys
  3. To define accountability, policies and standards for data quality, security and use across the organisation Correct answer.
  4. To speed up the execution of SQL queries on the warehouse

Explanation

Data governance assigns roles such as data owners and stewards and sets policies for quality, privacy, security and permitted use. It is an organisational control framework rather than a performance-tuning or procurement activity.

Question 18Data Analytics Fundamentals

Under the GDPR, which principle requires that personal data be collected for specified, explicit and legitimate purposes and not reused incompatibly?

  1. Integrity and confidentiality
  2. Purpose limitation Correct answer.
  3. Storage limitation
  4. Data minimisation

Explanation

Purpose limitation restricts processing to the declared, legitimate purposes for which data was collected. Storage limitation concerns retention periods, minimisation concerns how much data is collected, and integrity covers security safeguards.

Question 19Data Analytics Fundamentals

A dashboard shows monthly revenue as a bar chart, but the vertical axis starts at 90,000 instead of zero. What is the main risk of this design choice?

  1. Small differences between months appear far larger than they are Correct answer.
  2. The underlying data are automatically corrupted
  3. The chart becomes impossible to render in most tools
  4. The chart can no longer show more than two categories

Explanation

A truncated axis compresses the visible range, so modest differences are visually exaggerated and can mislead viewers. Bar charts in particular should normally start at zero because bar length encodes magnitude.

Question 20Data Analytics Fundamentals

Which chart type is best suited to showing how a single metric evolved over consecutive months?

  1. Pie chart
  2. Stacked bar chart with unrelated categories
  3. Line chart Correct answer.
  4. Scatter plot

Explanation

A line chart places time on one axis and connects successive values, making trends and turning points easy to read. Pie charts show composition at one moment, and scatter plots relate two numeric variables.

Question 21Data Analytics Fundamentals

What does ETL stand for in the context of data integration?

  1. Export, Transfer, Log
  2. Extract, Transform, Load Correct answer.
  3. Evaluate, Test, Launch
  4. Explore, Train, Learn

Explanation

ETL describes extracting data from sources, transforming it into a consistent and clean form, and loading it into a target store such as a warehouse. ELT reverses the order by loading raw data first and transforming it inside the target system.

Question 22Data Analytics Fundamentals

An analyst needs to count orders per region from a table with columns order_id, region and amount. Which SQL clause combination produces that result?

  1. SELECT region, COUNT(*) FROM orders GROUP BY region Correct answer.
  2. SELECT region FROM orders ORDER BY amount
  3. SELECT DISTINCT amount FROM orders
  4. SELECT COUNT(*) FROM orders WHERE region = region

Explanation

Grouping by region collapses rows into one per region, and COUNT(*) then returns the number of orders in each group. Aggregate functions paired with GROUP BY are the standard way to produce per-category totals.

Question 23Data Analytics Fundamentals

A data steward finds that the same customer appears twice with slightly different spellings of the name and one shared email address. Which data quality dimension is primarily violated?

  1. Granularity
  2. Uniqueness Correct answer.
  3. Timeliness
  4. Accessibility

Explanation

Duplicate records describing one real-world entity violate uniqueness, which requires each entity to appear once. Timeliness concerns currency of data, accessibility concerns availability, and granularity concerns the level of detail captured.

Question 24Data Analytics Fundamentals

In a decision tree, what does the root node represent?

  1. A random feature selected without any criterion
  2. A leaf holding the model's error rate
  3. The final predicted class for a record
  4. The first split chosen because it best separates the target classes Correct answer.

Explanation

The root node is the topmost split, selected by an impurity criterion such as Gini or information gain to best separate the classes. Predictions are produced at the leaf nodes reached after following the splits.

Question 25Data Analytics Fundamentals

A data lake differs from a traditional data warehouse mainly in that a data lake:

  1. Enforces a strict relational schema at the moment of ingestion
  2. Cannot be queried by analytics tools
  3. Stores raw data in its native format, often including unstructured and semi-structured content Correct answer.
  4. Only accepts data that has already been aggregated

Explanation

A data lake typically ingests raw files in native formats and applies structure later, on read, which suits diverse and unstructured sources. A warehouse usually applies schema-on-write with a predefined relational model.

Question 26Data Analytics Fundamentals

Which statement best describes the difference between data and information within the field of data analytics?

  1. Data refers to the results of an analysis, while information refers to the input of that analysis.
  2. Data always consists of numbers, while information always consists of text and images.
  3. Data consists of raw, unprocessed facts, while information is data that has been organised and given context so it becomes meaningful. Correct answer.
  4. Data is stored in databases, while information exists only in printed reports.

Explanation

In data analytics, data are raw recorded facts such as values in a table, and they gain meaning only when processed and placed in context. The processed, contextualised result is called information, which supports interpretation and decisions.

Question 27Data Analytics Fundamentals

An analytics team classifies its work into descriptive, diagnostic, predictive and prescriptive analytics. A report explaining why sales dropped in a specific region last quarter belongs to which category?

  1. Diagnostic analytics Correct answer.
  2. Descriptive analytics
  3. Prescriptive analytics
  4. Predictive analytics

Explanation

Diagnostic analytics goes beyond summarising what happened and investigates the causes behind an observed outcome, such as drilling into the drivers of a regional sales decline. Descriptive analytics would only report the drop, while predictive and prescriptive analytics look forward.

Question 28Data Analytics Fundamentals

Which characteristic is typical of structured data as handled in a relational analytics environment?

  1. It is organised into a predefined schema of rows and columns with defined data types. Correct answer.
  2. It is stored only as free-form text documents without any schema.
  3. It cannot be queried with SQL because it has no fixed fields.
  4. It must always be captured as video or audio streams.

Explanation

Structured data conforms to a predefined model, typically tables with named columns and declared data types, which makes it easy to query with SQL. Free-form text, images and audio are usually classed as semi-structured or unstructured.

Question 29Data Analytics Fundamentals

A retail analyst records each product with a code such as 'SKU-1042'. Which level of measurement does this product code represent?

  1. Nominal Correct answer.
  2. Ordinal
  3. Ratio
  4. Interval

Explanation

A product code is a label that identifies a category; the digits carry no meaningful order or arithmetic distance, so the scale is nominal. Ordinal, interval and ratio scales all require some ordering or numeric meaning that an identifier code does not have.

Question 30Data Analytics Fundamentals

Which statement correctly describes the role of a data dictionary in analytics work?

  1. It defines the visual layout of dashboards and reports.
  2. It documents the meaning, format and allowed values of each data element so that users interpret fields consistently. Correct answer.
  3. It stores a copy of every transaction processed by the operational systems.
  4. It automatically corrects errors in the source data before loading.

Explanation

A data dictionary is metadata that describes each field: its definition, data type, format, allowed values and ownership. This shared reference prevents different analysts from interpreting the same column in conflicting ways.

Question 31Data Analytics Fundamentals

An analyst wants a quick visual summary of the distribution of a single numeric variable, including its median and outliers. Which technique is most appropriate?

  1. A stacked bar chart of the variable
  2. A pie chart of the variable
  3. A scatter plot of the variable against itself
  4. A box plot of the variable Correct answer.

Explanation

A box plot displays the median, quartiles, spread and potential outliers of one numeric variable in a compact form. Pie charts and stacked bars are for part-to-whole comparisons of categories, and a scatter plot needs two different variables.

Question 32Data Analytics Fundamentals

In the CRISP-DM reference model, which task belongs to the data understanding phase rather than to data preparation?

  1. Selecting and cleaning records for modelling
  2. Collecting initial data and describing its basic statistical properties Correct answer.
  3. Transforming and scaling variables into model-ready form
  4. Constructing derived attributes from existing fields

Explanation

Data understanding covers initial data collection, description and exploration of basic properties, such as counts and summary statistics. Cleaning, transformation, scaling and deriving new attributes are all data preparation activities.

Question 33Data Analytics Fundamentals

An analyst computes the mean, median and mode of a survey variable. Which term describes these summary values collectively?

  1. Measures of association
  2. Measures of central tendency Correct answer.
  3. Measures of shape
  4. Measures of dispersion

Explanation

Mean, median and mode all describe the centre or typical value of a distribution, so they are measures of central tendency. Dispersion covers spread, association covers relationships between variables, and shape covers skewness and kurtosis.

Question 34Data Analytics Fundamentals

A sample of delivery times has a standard deviation of 4 minutes. What does this value express?

  1. The difference between the longest and shortest delivery time
  2. The average of the squared delivery times
  3. The typical distance of individual delivery times from the sample mean Correct answer.
  4. The most frequently occurring delivery time

Explanation

The standard deviation is the square root of the variance and expresses typical deviation of observations from the mean in the original units. The range is the max-min difference, the mode is the most frequent value, and the mean of squares is a computational step, not the deviation itself.

Question 35Data Analytics Fundamentals

In a probability distribution, what does the term 'expected value' refer to?

  1. The largest value the variable can take
  2. The long-run average outcome if the random process were repeated many times Correct answer.
  3. The spread of outcomes around the median
  4. The outcome with the highest single probability

Explanation

The expected value is the probability-weighted average of all possible outcomes, interpreted as the long-run mean over repeated trials. It is not necessarily the most likely single outcome nor an extreme value of the variable.

Question 36Data Analytics Fundamentals

A quality team measures the diameter of manufactured bolts, and the measurements cluster around a target with roughly equal spread on both sides. Which distribution is the standard model for this continuous measurement?

  1. The binomial distribution
  2. The uniform distribution over integers
  3. The normal distribution Correct answer.
  4. The Poisson distribution

Explanation

Continuous measurements that vary symmetrically around a central target are commonly modelled with the normal distribution. The binomial and Poisson distributions model counts of discrete events, and a discrete uniform distribution assigns equal probability to a finite set of integers.

Question 37Data Analytics Fundamentals

A marketing analyst wants to test whether average basket value differs between two independent customer groups. Which family of statistical methods is designed for this purpose?

  1. Hypothesis testing comparing group means Correct answer.
  2. Data lineage tracking
  3. Cluster analysis
  4. Time-series decomposition

Explanation

Comparing the means of two independent groups is a hypothesis-testing problem, typically addressed with a two-sample t-test or a nonparametric equivalent. The other options concern forecasting, unsupervised grouping and metadata tracking, none of which tests a difference in means.

Question 38Data Analytics Fundamentals

An analyst reports a 95% confidence interval for average order value as 48 to 56 euros. What does the 95% refer to?

  1. The percentage of sample observations that fall between 48 and 56 euros
  2. The proportion of intervals built this way that would contain the true population mean Correct answer.
  3. The probability that the true mean lies between 48 and 56 euros
  4. The chance that a future sample mean equals the true population mean

Explanation

The confidence level describes the long-run performance of the procedure: across repeated samples, about 95% of intervals constructed this way would capture the true parameter. It is not the probability that this particular interval contains the fixed true mean, nor a statement about individual observations.

Question 39Data Analytics Fundamentals

Which task is a core responsibility of a data analyst rather than of a database administrator?

  1. Managing backup schedules and storage allocation for database servers
  2. Configuring user access rights on the database management system
  3. Applying operating system patches to database servers
  4. Interpreting data to answer business questions and communicating findings to stakeholders Correct answer.

Explanation

The data analyst's role centres on querying, interpreting and presenting data to support decisions. Backups, storage, access configuration and patching are operational administration duties that keep the database platform running.

Question 40Data Analytics Fundamentals

A firm introduces a policy that customer email addresses may only be used for the purpose for which they were collected. Which data protection principle does this policy implement?

  1. Data minimisation
  2. Purpose limitation Correct answer.
  3. Storage limitation
  4. Integrity and confidentiality

Explanation

Purpose limitation requires personal data to be collected for specified, explicit and legitimate purposes and not reused for incompatible ones. Data minimisation concerns collecting only what is necessary, storage limitation concerns retention periods, and integrity and confidentiality concern secure handling.

Question 41Data Analytics Fundamentals

What is the primary purpose of establishing a data quality dimension such as 'completeness'?

  1. To guarantee that every database column contains only numeric values
  2. To determine how long records must be retained
  3. To define a measurable criterion for judging whether required data values are present Correct answer.
  4. To specify which employees may view the data

Explanation

Data quality dimensions provide measurable criteria for assessing data fitness for use, and completeness specifically measures whether required values are present. It does not impose data types, control access or set retention rules.

Question 42Data Analytics Fundamentals

An analyst must merge two datasets that use different formats for the same date field, one as 'DD/MM/YYYY' and the other as 'YYYY-MM-DD'. Which preparation step resolves this before analysis?

  1. Standardising the date format through transformation Correct answer.
  2. Aggregating both datasets by month before merging
  3. Removing all rows containing dates
  4. Encrypting the date column in both datasets

Explanation

Data preparation includes transformation steps that harmonise formats so records can be joined and compared correctly. Standardising both fields to one representation preserves the data while removing the mismatch.

Question 43Data Analytics Fundamentals

A dataset of website visits records the number of pages viewed per session. Which data type best describes this variable?

  1. Nominal categorical
  2. Continuous numerical
  3. Discrete numerical Correct answer.
  4. Ordinal categorical

Explanation

Page views are counted in whole units, so the variable is discrete numerical. It is not continuous because fractional page views are impossible, and it is not categorical because the values support arithmetic operations such as totals and averages.

Question 44Data Analytics Fundamentals

An analyst samples every tenth record from a list of customers sorted by registration date. Which sampling method is being used?

  1. Cluster sampling
  2. Systematic sampling Correct answer.
  3. Simple random sampling
  4. Stratified sampling

Explanation

Selecting every k-th element from an ordered list is systematic sampling, which is easy to apply but can be biased if the list has a periodic pattern. Simple random sampling gives every record an equal independent chance, while stratified and cluster sampling divide the population into groups first.

Question 45Data Analytics Fundamentals

An analysis of customer churn reports that customers with a support ticket in the last month churn at twice the rate of those without. Which type of analytics output does this represent?

  1. A causal proof that support tickets cause churn
  2. A forecast of next year's churn rate
  3. A descriptive insight about an observed association Correct answer.
  4. A recommendation of which retention campaign to run

Explanation

The finding describes an association observed in historical data and does not by itself establish causation or predict the future. Establishing causality would require a designed study or causal model, and forecasting and recommendation are separate analytical tasks.

Question 46Data Analytics Fundamentals

Which statement correctly distinguishes population from sample in statistical analysis?

  1. The population is the subset measured, while the sample is the entire set of units of interest.
  2. The population refers only to people, while the sample refers only to numeric records.
  3. The population is always larger than one million units, while the sample is always smaller than one thousand.
  4. The population is the entire set of units of interest, while the sample is a subset selected for measurement. Correct answer.

Explanation

The population is the complete group about which conclusions are wanted, and the sample is the subset actually observed. Inferential statistics use sample results to draw conclusions about the population.

Question 47Data Analytics Fundamentals

An analyst notices that a sales dataset has a strong seasonal pattern repeating every twelve months. Which analytical approach is specifically designed to handle this structure?

  1. Time-series analysis with seasonal decomposition Correct answer.
  2. Cluster analysis of product categories
  3. A chi-square test of independence
  4. Cross-sectional regression on a single period

Explanation

Time-series methods model observations ordered in time and can separate trend, seasonal and residual components, which is exactly what a repeating twelve-month pattern requires. Cross-sectional, clustering and independence tests ignore the temporal ordering.

Question 48Data Analytics Fundamentals

A dashboard is intended for executives who need a quick overview of key performance indicators. Which design choice best supports that purpose?

  1. Use many colours and 3D effects to make the dashboard visually striking
  2. Include every available field so no information is omitted
  3. Place raw transaction-level tables on the first screen
  4. Show a small set of clearly labelled headline indicators with minimal detail Correct answer.

Explanation

Executive dashboards work best when they highlight a few key indicators with clear labels and leave detail available on demand. Overloading the screen with fields, decorative effects or raw records obscures the message and slows interpretation.

Question 49Data Analytics Fundamentals

Which situation best illustrates the concept of a data silo within an organisation?

  1. A data warehouse consolidates records from several source systems into a single integrated repository.
  2. The marketing department keeps its customer data in a system that other departments cannot access or combine with their own. Correct answer.
  3. An analyst exports a report to a spreadsheet for a one-off presentation.
  4. A team publishes a shared data dictionary that all departments use.

Explanation

A data silo exists when a group's data is isolated so that others cannot access or integrate it, which fragments the organisational view. Consolidation in a warehouse, shared dictionaries and one-off exports do not describe such isolation.

Question 50Data Analytics Fundamentals

An analyst wants to compare the distribution of salaries across four departments on a single figure. Which visualisation is most suitable?

  1. A single pie chart with four slices
  2. A line chart with departments on the horizontal axis
  3. A scatter plot of salary against department name
  4. Side-by-side box plots, one per department Correct answer.

Explanation

Side-by-side box plots place the distributions of several groups on a common scale, allowing medians, spread and outliers to be compared directly. Pie charts show parts of a whole, and line or scatter plots are poorly suited to comparing distributions across categories.