Study guide
8.1 Data Representation and Analysis
National Benchmark Tests NBT, South Africa
Track your progress as you read
Finished sections get a check mark.
1 Introduction & Exam Overview
1.1 Current Exam Facts 1.2 Who Writes Which Test 1.3 Format, Timing, Languages & Materials
2 Booking, Fees & Results
2.1 Registration, Venues, Remote Writers & Disabilities 2.2 Fees, Payment & Retakes 2.3 Benchmarks, Results & University Use
3 Academic Literacy: Meaning
3.1 Communicative Function 3.2 Inferencing 3.3 Vocabulary in Context 3.4 Grammar and Syntax
4 Academic Literacy: Structure
4.1 Cohesion 4.2 Discourse and Argument 4.3 Essential versus Non-Essential Information
5 Academic Literacy: Genre and Visuals
5.1 Metaphor and Idiom 5.2 Text Genre, Audience, Purpose and Tone 5.3 Visual and Numerical Information in Text
6 Quantitative Literacy: Number
6.1 Quantity, Number and Operations 6.2 Ratios and Relative Difference 6.3 Multi-Step Calculation without a Calculator
7 Quantitative Literacy: Pattern and Space
7.1 Relationships, Patterns and Simple Algebra 7.2 Change and Rates 7.3 Shape, Dimension and Space
8 Quantitative Literacy: Data
8.1 Data Representation and Analysis 8.2 Chance and Uncertainty 8.3 Integrating Tables, Graphs and Text
9 MAT: Numbers and Algebra
9.1 Real Numbers, Surds and Exponents 9.2 Sequences and Series 9.3 Algebraic Manipulation, Equations and Inequalities
10 MAT: Modelling, Logs and Finance
10.1 Ratios, Percentages and Modelling 10.2 Logarithms, Exponential Equations and Financial Mathematics 10.3 Mathematical Modelling and Process Skills
11 MAT: Functions
11.1 Linear and Quadratic Functions 11.2 Hyperbola, Exponential and Logarithmic Functions 11.3 Graphs, Equations and Inequality Regions 11.4 Differential Calculus: Gradients and Turning Points
12 MAT: Transformations and Trigonometry I
12.1 Transformations of Functions 12.2 Trigonometric Ratios and Graphs 12.3 Trigonometric Identities and Equations
13 MAT: Trigonometry II and Space
13.1 Trigonometry in Two and Three Dimensions 13.2 Shapes, Perimeter, Area and Volume 13.3 Analytic Geometry
14 MAT: Circle Geometry, Data and Logic
14.1 Circle Geometry 14.2 Data Handling and Probability 14.3 Logical Deductions and Validity
15 Official Prep and Session Strategy
15.1 Official Prep, Exemplars and NBT versus NSC 15.2 AQL Session Strategy 15.3 MAT Same-Day Strategy
8.1 Data Representation and Analysis
Key Takeaways Explain This Topic AI
NBT QL data items ask you to decode a table or chart and then interpret it in a higher-education context, not only to read an axis label
Mixed tables combine counts, rates, and categories; the largest occupied-bed count is not automatically the highest occupancy rate
Pie charts show parts of one whole; grouped bars compare side by side; stacked bars show composition; broken lines show change over time; scatter plots show association
The mean follows outliers, the median is the middle value, and the mode is the most frequent value; choose the measure that matches the story
Tree branches must add back to their parent total; a leftover that does not match the title means you are reading a subset or an incomplete split
Last updated: September 2026
Decode the display, then interpret it
The Quantitative Literacy (QL) component of South Africa's National Benchmark Tests sits inside the Academic and Quantitative Literacy (AQL) paper. It is not a second Mathematics (MAT) paper. The National Benchmark Tests Project, through the Centre for Educational Assessments, describes data representation and analysis as the ability to derive information from contextualised tables, charts and diagrams and then to interpret what that information means. Independent ExamFinch study material for NBT QL trains that as a two-step habit: decode the display, then say what the numbers imply for a real campus decision.
QL items use tables with several rows and columns and mixed data types. A single table may mix whole-number counts, percentages, categories, and time labels. The trap is to treat every column as if it answered the same question.
Consider a fictional first-year residence snapshot at Marula University, a public university in Limpopo, at the start of the academic year.
- Residence Beds Occupied week 1 Occupied week 6 NSFAS-funded Self-funded Other funding Mean wait (days)
- Magoleng 240 228 236 174 54 8 11
- Baobab 180 162 179 121 41 17 18
- Mopani 320 288 301 240 48 13 9
- Ivory 90 71 88 22 49 17 4
Week 1 occupancy rates are 228/240 = 95% at Magoleng, 162/180 = 90% at Baobab, 288/320 = 90% at Mopani, and 71/90 ≈ 79% at Ivory. Mopani housed the most students (288), but Magoleng filled the largest share of its beds. Ivory looks "empty" if you only count vacant beds (19), yet Mopani had 32 vacant beds — a larger count of empty rooms and still a higher fill rate, because Mopani is much bigger. On QL, "highest occupancy" is incomplete until you know whether the setter wants a count or a rate.
Check that mixed columns are consistent. NSFAS-funded + self-funded + other funding at Magoleng is 174 + 54 + 8 = 236, which matches week 6 occupancy, not week 1 and not the bed total. If a question asks for the NSFAS share of Magoleng residents in week 6, the denominator is 236 occupied beds, giving 174/236 ≈ 73.7%, not 174/240. Using the design capacity as the base quietly changes the story.
Pie charts show parts of one whole. A pie of Magoleng's week 6 funding mix is appropriate because NSFAS, self-funded, and other are mutually exclusive categories that add to the occupied total. A pie is a poor choice for comparing Magoleng with Ivory, or for showing change from week 1 to week 6: those are not slices of a single whole. If someone draws two pies, one per week, you still cannot read a trend off a slice angle without converting back to counts or rates. QL often asks whether a pie is even a fair display, not only which slice is largest.
Simple bar charts compare separate categories with a common unit. Grouped (compound) bar charts place two or more bars side by side for each category — for example NSFAS versus self-funded counts at each residence — so you can compare funding mix across residences without forcing the bars to add to a total you care about. Stacked bars pile segments that do add to a total: NSFAS, self-funded, and other stacked to week 6 occupancy. Read a stacked bar from the baseline for the bottom segment only. The middle segment's length is its own count; its starting height is not its value. Candidates who read a stacked mid-segment as if it started at zero invent a number the chart never drew.
A broken-line graph joins successive time points. Marula's campus clinic logged influenza-like visits in the first eight teaching weeks as 18, 22, 31, 40, 38, 27, 21, 19. The line rises to week 4, then falls. "Broken" here means the graph is a polyline through those points, not a smooth curve you should interpolate as if visits were continuous between Wednesdays. You may estimate a midpoint only if the question asks for an approximate reading; you may not invent a week-3.5 value and treat it as measured. The interpretation is seasonal: visits peaked after registration week, then declined — not "the clinic is steadily busier."
Scatter plots show paired measurements for the same students. A first-year chemistry practical at Marula plotted tutorial sessions attended (horizontal) against practical mark out of 100 (vertical). Most points rise from left to right: students who attended more tutorials tended to score higher. That is an association, not a proof that tutorials cause the marks. One student attended 11 of 12 tutorials and scored 38. That point is an outlier relative to the cloud. It does not erase the overall positive pattern, and it does not prove that tutorials harm performance. On QL, name the outlier, then still describe the main trend.
Tree diagrams split a group into successive branches. Marula's 400 first-years split by faculty (Health Sciences 120, Education 80, Commerce 200), then each faculty splits into campus-shuttle users versus walk-or-taxi. Every child branch must add back to its parent: 48 shuttle + 72 walk = 120 Health students. If a tree's last row does not sum to the title total, the diagram is incomplete or you are reading a subset. Trees on QL are often classification devices, not MAT-style probability engines.
Mean, median, and mode are tools for a story, not decorations. Seven practical marks were 48, 51, 53, 55, 56, 58 and 94. The mean is 415 ÷ 7 ≈ 59.3. The median, the middle value once ordered, is 55. There is no mode; every mark appears once. The 94 is an outlier. Remove it and the mean becomes 321 ÷ 6 = 53.5 (a drop of about 5.8 marks) while the median becomes (53 + 55) / 2 = 54 (a drop of 1). The mean followed the extreme score; the median barely moved. A second example: ten weekly library no-show penalties of 0, 0, 1, 1, 1, 1, 2, 3, 3, 12. The mode is 1 (most frequent), the median is 1, and the mean is 24/10 = 2.4. Reporting only the mean would make no-shows sound routine; the mode and median say the typical week is a single penalty, with one extraordinary week of 12.
National reports treat Quantitative Literacy as a demanding domain: the 2025-intake National Benchmark Tests report listed a full-cohort QL mean near 49 percent. That figure is overall QL, not a published per-subdomain cut-score. Treat data items as interpretation under time pressure, not giveaways. Work without a calculator: cancel factors, estimate, and compare rather than chasing extra decimals. Read the title, units, and footnotes before you touch an option. Then interpret: what would a residence manager, a clinic nurse, or a first-year actually do with this number?