Data interpretation: pie charts, line graphs and caselets

In short

Data interpretation (DI) is the skill of pulling exact numbers out of a chart or a block of prose and answering questions about shares, totals, growth, averages and ratios. A pie chart shows composition at one moment, a line graph shows movement over time, and a caselet gives the same data in sentences with no chart at all, so you must build the table yourself. It is tested in the DILR section of CAT and the Data Analysis sections of IBPS/SBI PO and RRB, in SSC CGL Tier II, in the 15-mark General Aptitude paper of GATE (all branches), and in almost every campus test — TCS NQT, Infosys, Wipro NTH, Accenture, Capgemini, Cognizant, AMCAT, eLitmus and CoCubes. On the job it is the same arithmetic an MIS executive, business analyst or management trainee does on a dashboard every morning. Question counts and marking schemes change from year to year, so confirm the current pattern from the official notification.

Explained

What a DI set is made of, and the four question families

• Every set has three parts: the data (chart or prose), the units (percent, ₹ lakh, ₹ crore, tonnes, thousands), and the questions. Misreading the unit line costs more marks than weak arithmetic.
• Read in this order: title → axis labels and units → scale → footnotes ("figures in ₹ crore", "excluding exports") → only then the questions.
• Family 1 - direct read: "What was the production in 2022?" One lookup.
• Family 2 - comparison: differences, ratios, "by what percent more", "how many times".
• Family 3 - derived: averages, totals, growth rates, shares of a new total, per-unit values.
• Family 4 - judgement: "which of the statements is true", "in how many years was X above the average", "cannot be determined". These need every option checked, so they are the slowest - attempt them last.
• A "cannot be determined" option is often the right answer when the set gives only percentages and no absolute total, or only one year's total for a multi-year percentage chart.

Pie charts

• A pie chart divides a circle of 360° among the parts of one whole. So 1% = 3.6°, and 1° = 1/3.6 % ≈ 0.2778%.
• Data may be given as percentages, as degrees, or as absolute values; convert everything to one form before starting.
• Finding the hidden total: if a sector of 84° equals ₹10.5 lakh, the total is 10.5 × 360/84 = ₹45 lakh. One sector plus its angle fixes the whole chart.
• Share versus absolute value: a sector's share can fall while its value rises, if the total grew. This is the single most exploited trap in the topic.
• Two or more pie charts: when the totals differ, you may never compare percentages directly. Convert each to absolute values first (value = share × its own total). A 20% slice of 1,200 and a 16% slice of 1,500 are both 240 - only multiplying tells you which is bigger.
• Combining two pies: the joint share is a *weighted* average, not the simple average of the two percentages.
• Nested / double pie: the outer ring breaks one sector of the inner ring further; the outer percentages are shares of that sector, not of the grand total - multiply down the chain.
• Percentage versus percentage point: a share rising from 15% to 18% is a rise of 3 percentage points but of 20 percent (3/15).
• Strengths: instant sense of composition and of which part dominates. Limits: useless for trends, unreliable beyond about 6-7 sectors, and two pies cannot be compared by eye. Modern reporting practice prefers a stacked bar or a table for the same data.

Line graphs

• The x-axis is almost always time (years, quarters, months); the y-axis is the measured quantity. The slope between two points is the rate of change; the steepest segment is the largest change per period, which is not necessarily the largest percentage change.
• Absolute rise versus percentage rise: a jump from 40 to 60 is +20 and +50%; from 180 to 210 is +30 but only +16.67%. Questions asking "maximum growth" without the word "percent" mean absolute increase - read the wording twice.
• Multiple lines = multiple series (revenue and expenditure, two companies, two products). The vertical gap between them is the derived quantity (profit, deficit, lead).
• The intersection of two lines means the two quantities are equal at that point, not that either is maximum or zero.
• Rising line that is flattening still means growth, only slower; a line falling towards zero still means a positive quantity until it crosses the axis.
• Index lines: CPI, WPI, IIP and similar official series are published with a base year equal to 100 (for example the Indian CPI series uses 2024 = 100 since the January 2026 index, replacing 2012 = 100; the IIP and WPI series use 2022-23 = 100 since mid-2026, replacing 2011-12 - bases are revised from time to time). A value of 130 means 30% above the base year, not 130 units.
• Cumulative versus incremental: a cumulative line can never fall; if the plotted line falls, the data is period-wise.
• Reading traps: a y-axis that does not start at zero or has a break exaggerates the slope; a secondary (right-hand) y-axis means the two lines are on different scales; a dashed segment usually marks projected or provisional data.
• Strengths: trend, turning points, comparison of growth paths. Limits: poor for composition; too many lines becomes unreadable; interpolation between plotted points is an assumption, not data.

Caselets

• A caselet is DI in prose: three to six sentences of interlinked facts, then four to five questions. There is no chart, so the first move is always to draw the table or grid yourself before reading any question.
• Standard structures: a funnel (applied → shortlisted → tested → selected), a two-way grid (male/female × department, or city × product), a ratio chain (team split 7:5:4, then each split by gender), and a Venn caselet (students taking two or three subjects).
• Track the base of every percentage. "40% of those who appeared" is a percentage of a sub-group, not of the original total. Write each percentage as "x% of (what)" in the margin.
• When the total is not given, let the total be a convenient number - 100, or the LCM of all denominators appearing (for 3/4, 2/5 and 5/6, use 60) - solve, then scale to the real total at the end.
• When a quantity is unknown and relationships are linear, put x on it and form one equation; do not guess.
• Work sequentially down the funnel and write every intermediate number; most questions ask for an intermediate stage, not the final one.
• Venn caselets: use the inclusion-exclusion rule rather than trying to fill the diagram by trial.

Speed and accuracy technique

• Memorise fraction-percentage equivalents - they turn multiplications into divisions.
• Compute percentage change always on the old value (the base), never on the new one.
• For a ratio, change the numerator and denominator separately: numerator +20% with denominator +10% gives 1.20/1.10 − 1 ≈ +9.09%.
• Compare two fractions by cross-multiplication instead of converting both to decimals: 17/23 vs 22/31 → 17×31 = 527, 22×23 = 506, so 17/23 is larger.
• Approximate deliberately when options are far apart (round 2,987 to 3,000), but compute exactly when two options are within 2-3% of each other.
• Average of percentages is not the percentage of totals. Overall margin = total profit / total revenue; the mean of yearly margins is a different number.
• Use CAGR when the question asks for compounded annual growth over the period; use a simple annualised rate (total percentage change ÷ number of years) when it asks for growth without compounding; the mean of the yearly growth rates is a third, different figure - answer the one the wording names.
• Set selection matters: scan all sets for 20-30 seconds, do the one with the cleanest numbers first, and leave any set whose data you cannot read in two passes.
• Factor in negative marking - in most bank and SSC objective papers a wrong answer costs 0.25 of the mark, and in GATE MCQs 1/3 or 2/3 of the mark, with no penalty on numerical-answer questions; schemes vary, so check the current brochure.

Key points

• 360° = the whole - 1% of a pie equals 3.6°; angle = share% × 3.6; share% = angle ÷ 3.6.
• Sector value = (angle/360) × total = share% × total; the total itself = value × 360 ÷ its angle.
• Share ≠ value - a slice can shrink in percentage and still grow in rupees when the total expands.
• Two pies, two totals - convert to absolute numbers before any comparison; percentages across different wholes are not comparable.
• Combined share is a weighted average of the two shares, weighted by the two totals.
• Percentage point is the arithmetic difference of two percentages; percent change divides that difference by the original.
• Percentage change base is always the earlier or reference value, so a fall of 20% followed by a rise of 20% does not return to the start.
• Successive change of a% then b% equals a + b + ab/100 percent overall.
• Slope on a line graph is the rate of change; steepest segment = largest absolute change, not largest percentage change.
• Intersection of two lines means equality of the two quantities at that instant, nothing more.
• Index values are relative to a base year taken as 100; growth between two index values is their ratio minus one.
• CAGR smooths an uneven path into one compounded annual rate and never exceeds the simple average of the yearly growth rates - it is strictly below it whenever the yearly rates differ.
• Caselet first move - build the table; second move - label the base of every percentage.
• Assume the total as 100 or an LCM when a caselet gives only fractions and percentages.
• Overall ratio ≠ average of ratios - total numerator over total denominator is the only correct aggregate.
• "Cannot be determined" is a genuine answer when only shares are given and the absolute totals are missing.
• Check the unit line and the footnotes - lakh versus crore, thousands, "provisional", "excluding exports".
• Broken or non-zero y-axis exaggerates a trend; read values, not visual steepness.

Formulas and examples

Example 1 - pie chart given in degrees. A panchayat's annual budget of ₹45 lakh is divided as: Roads 72°, Water 54°, Health 90°, Education 60°, Sanitation 84°. (Check: 72 + 54 + 90 + 60 + 84 = 360°.)

• Value per degree = 45,00,000/360 = ₹12,500.
• Roads = 72 × 12,500 = ₹9,00,000; Water = ₹6,75,000; Health = ₹11,25,000; Education = ₹7,50,000; Sanitation = ₹10,50,000 (sum = ₹45 lakh ✓).
• Health as a share of the budget = 90/3.6 = 25%.
• Health exceeds Water by (11.25 − 6.75)/6.75 × 100 = 4.5/6.75 = 66.67%.
• If next year Education is to get ₹9,00,000 out of the same ₹45 lakh, its angle = (9/45) × 360 = 72°, a rise of 12° or 3.33 percentage points of share.

Example 2 - two pie charts, different totals. College A has 1,200 final-year students, College B has 1,500. Stream shares: A - CSE 30%, ECE 25%, Mech 20%, Civil 15%, Others 10%; B - CSE 36%, ECE 20%, Mech 16%, Civil 18%, Others 10%.

• A: CSE 360, ECE 300, Mech 240, Civil 180, Others 120 (= 1,200 ✓). B: CSE 540, ECE 300, Mech 240, Civil 270, Others 150 (= 1,500 ✓).
• Ratio of CSE students, A : B = 360 : 540 = 2 : 3; B's CSE exceeds A's by 180, i.e. 50%, though the share gap is only 6 percentage points.
• Mech: 20% of 1,200 and 16% of 1,500 both equal 240 - the larger percentage does not mean more students.
• ECE is likewise equal at 300 each.
• Civil, both colleges together = 180 + 270 = 450 out of 2,700 = 16.67%.
• Combined CSE share = 900/2,700 = 33.33%, not the simple average (30 + 36)/2 = 33% - it is weighted by the two totals.

Example 3 - line graph with two series. Revenue and expenditure of a firm (₹ crore):

• Maximum profit is 2023 (₹50 crore), but maximum profit percent is 2020 (30%) - two different answers from the same graph.
• Year of highest revenue growth: 2020 = 36/120 = 30.00%; 2021 = 24/156 = 15.38%; 2022 = 12/180 = 6.67%; 2023 = 58/192 = 30.21% - so 2023, by a margin no eye could judge. Compute, do not eyeball.
• Average annual profit = (20 + 36 + 36 + 32 + 50)/5 = 174/5 = ₹34.8 crore.
• Overall margin on revenue for the five years = 174/898 = 19.38%, while the mean of the five yearly margins on revenue is a different figure - the two are not interchangeable.
• CAGR of revenue, 2019 to 2023 (n = 4 years) = (250/120)^1/4 − 1 = (2.0833)^0.25 − 1 ≈ 1.2014 − 1 = about 20.1% per annum. Verification: 120 × 1.2014⁴ ≈ 250 ✓.

Example 4 - caselet, placement funnel. 600 final-year students appeared for a campus drive. 60% cleared the aptitude test. Of those who cleared, three-fourths appeared for the technical round. 40% of those who appeared failed the technical round. Every student who cleared the technical round attended the HR round, and 5/6 of them received offers.

• Cleared aptitude = 0.60 × 600 = 360.
• Appeared for technical = (3/4) × 360 = 270; did not appear = 90.
• Cleared technical = 60% of 270 = 162; failed = 108.
• Offers = (5/6) × 162 = 135.
• Offers as a percentage of all who appeared = 135/600 = 22.5%.
• Cleared aptitude but got no offer = 360 − 135 = 225.
• Note the shifting base: 40% applied to 270 (not to 360 or 600) and 5/6 applied to 162.

Example 5 - caselet, ratio plus percentage grid. A firm's 480 employees are engineers, sales and support staff in the ratio 7 : 5 : 4. Women are 40% of engineers, 60% of sales and 50% of support.

• Parts total 7 + 5 + 4 = 16, so one part = 480/16 = 30 → Engineers 210, Sales 150, Support 120 (= 480 ✓).
• Women: 0.40 × 210 = 84; 0.60 × 150 = 90; 0.50 × 120 = 60 → total 234 women, i.e. 234/480 = 48.75% of staff - again a weighted average of 40%, 60% and 50%, not their mean (50%).
• Men among engineers = 126; ratio of male engineers to female sales staff = 126 : 90 = 7 : 5.
• If 20 women are hired into sales and no one leaves, new total = 500, women = 254 → 50.8%.

Where it is used

• Campus aptitude tests and government exams: DI sets in TCS NQT, Infosys, Wipro NTH, Cognizant and Accenture tests; the Data Analysis & Interpretation section of IBPS/SBI PO mains; SSC CGL Tier II; RRB NTPC; the General Aptitude section common to all GATE papers.
• MIS and business-analyst roles: daily and monthly dashboards in Excel, Power BI or Tableau - share of revenue by region, month-on-month growth, funnel conversion - the identical arithmetic done on live data.
• Banking, insurance and NBFC operations: branch-wise deposit mix (a pie), NPA trend over quarters (a line), and loan-sanction funnels (a caselet) in management-trainee and PO work.
• Market research and FMCG sales: Nielsen-style market-share pies, monthly offtake lines, and distributor coverage caselets at firms such as HUL, ITC and Dabur.
• Public data and policy work: reading RBI bulletins, MoSPI/NSO releases, the Economic Survey, Census and NFHS tables - sectoral GDP pies, CPI and IIP index lines, and footnote-heavy survey caselets.
• Operations and BPO/WFM analytics: call-volume and AHT trend lines, shrinkage composition pies, and staffing caselets used in capacity planning.

Interview questions

• Q: What is the difference between a percentage change and a percentage point change? A: A share moving from 15% to 18% is a 3 percentage point rise but a 20% rise, because percentage change divides the 3 by the original 15. Interviewers use this to test precision in reporting.
• Q: Can a product's market share fall while its sales rise? A: Yes - if the whole market grew faster than the product did. Share is relative to the total, so share and absolute value can move in opposite directions.
• Q: Two pie charts show 20% and 16% for the same category. Which has more? A: Cannot be said without the two totals; 20% of 1,200 and 16% of 1,500 are both 240.
• Q: When would you use a pie chart and when a line graph? A: A pie for the composition of one whole at one point in time with few categories; a line for movement of a quantity over time or for comparing growth paths. For composition over time, a stacked bar or multiple lines is better than several pies.
• Q: What does the point where two lines cross mean? A: Only that the two quantities are equal there - it is not a maximum, a minimum or a zero.
• Q: Why is CAGR preferred to the average of yearly growth rates? A: CAGR is the single compounded rate that actually takes the initial value to the final value; averaging yearly rates ignores compounding and overstates growth when the series is volatile.
• Q: A caselet gives only percentages and fractions, no totals. How do you proceed? A: Assume the total as 100 or as the LCM of all denominators, solve completely, then scale if a real total appears later.
• Q: How do you spot a misleading chart? A: A y-axis not starting at zero or with a break, a dual axis with different scales, a 3-D pie that distorts slice sizes, truncated time ranges, and missing or hidden units and footnotes.
• Q: The company's revenue index moved from 100 to 144 over four years. What was the annual growth? A: CAGR = 1.44^1/4 − 1 ≈ 9.54% per annum - not 11%, which is the simple average of the 44% total.
• Q: If the numerator of a ratio rises 25% and the denominator rises 25%, what happens to the ratio? A: Nothing - the ratio is unchanged, because both scale by the same factor.
• Q: How do you manage time across four DI sets with only 15 minutes? A: Scan all four for about half a minute, pick the sets with round numbers and direct questions, finish those fully, and leave judgement-type "which statements are true" questions for the end since each needs every option checked.
• Q: A sector is 54°. What percentage and what value does it represent if the whole is ₹80 lakh? A: 54/3.6 = 15%, so 0.15 × 80 = ₹12 lakh.

Common mistakes

• Comparing percentages across two pie charts with different totals instead of converting both to absolute values first.
• Taking the percentage change on the new value rather than on the original base, or reporting a percentage point gap as a percentage rise.
• Answering "maximum growth" with the largest percentage rise when the question asked for the largest increase, or the reverse - the wording decides.
• Averaging ratios or percentages instead of dividing the total numerator by the total denominator (average of margins is not the overall margin).
• Losing track of the base in a caselet - applying a percentage to the original total when the sentence applied it to a sub-group.
• Missing the unit or footnote line: reading ₹ crore as ₹ lakh, treating "figures in thousands" as units, or ignoring a non-zero or broken y-axis and judging a trend by visual steepness.
• Over-approximating when two options are close, or conversely doing exact long division when the options are far apart and time is short.

Related topics

Percentages · Ratio, proportion and partnership · Averages, mixtures and alligation · Data interpretation · Simplification and approximation · Excel charts, pivot tables and dashboards.