Data Interpretation — CAT DILR

Data Interpretation questions arrive as sets: one table, chart or combination of both, followed by 4-6 questions that each extract a different fact from it. The mathematics involved — percentages, ratios, averages, growth rates — is exactly the Arithmetic chapter's toolkit. What DI actually tests is reading discipline: extracting the right numbers from the right rows and columns under time pressure, without re-reading the source data for every sub-question.

1. The DI mindset — read the questions before the data

Read all 4-6 questions in a set before studying the data table in detail. The questions tell you which rows, columns or years actually matter — a table might have 8 columns of data with only 3 of them ever referenced across the whole question set. Studying the entire table exhaustively before knowing what will be asked wastes the scarce 40-minute DILR window on information that may never be used.

Do one careful full read of the table's structure — what each row and column represents, and the units — before answering any individual question. This single read prevents the far more costly error of misreading a column header mid-set and getting several questions wrong from the same misunderstanding.

2. Percentage, percentage-point and percentage-change reading

$\%\text{ change}=\frac{\text{new}-\text{old}}{\text{old}}\times 100$

A "percentage point" change and a "percentage" change are different quantities read off the same data, and DI sets deliberately test which one a question asks for. If a company's market share moves from 20% to 25%, that is a rise of 5 percentage points, but a $\frac{25-20}{20}\times 100=25\%$ relative increase in market share itself.

A question asking "by what percentage did the share increase" wants the second number; a question asking "by how many percentage points" wants the first — reading the question's exact wording before computing avoids answering the wrong one correctly.

3. Ratios and share of total

When a table gives a "total" row or column, most questions about individual shares reduce to one division. The share of item $i$ in a total is $\frac{{\text{value}}_i}{\text{total}}\times 100$, and a "combined share of A and B" question is simply $\frac{{\text{value}}_A+{\text{value}}_B}{\text{total}}\times 100$ — no more complex than adding two numbers before dividing.

Trap. Pie charts showing percentage shares are especially prone to a rounding trap: individual segment percentages, each independently rounded, may sum to $99\%$ or $101\%$ rather than exactly $100\%$. A question requiring the exact underlying values (not just the rounded percentages shown) needs the actual total figure, which is sometimes given separately from the chart.

4. Averages and weighted averages

A simple average across a table's rows treats every row equally; a weighted average does not, and DI sets often supply the weights implicitly through a stated total that must be used instead of a naive row-count average. If a student's marks in three subjects are 80, 90 and 70, with those subjects weighted 2, 3 and 5 out of a total weight of 10, the weighted average is

$\frac{80\times 2+90\times 3+70\times 5}{10}=\frac{160+270+350}{10}=78$

not the simple average of $\frac{80+90+70}{3}=80$ — a full 2 marks higher than the correct weighted figure, because the naive average ignores that the lowest-scoring subject carries the most weight.

5. Growth rates and multi-year comparisons

Year-over-year percentage change and the compound annual growth rate (CAGR) answer genuinely different questions, and DI sets exploit the difference. Year-over-year change compares two adjacent years directly; CAGR smooths a multi-year change into a single constant annual rate:

$\text{CAGR}={\left(\frac{\text{final value}}{\text{initial value}}\right)}^{\frac{1}{n}}-1$

where $n$ is the number of years spanned. A company growing from ₹200 crore to ₹360 crore over 4 years has a CAGR of ${\left(\frac{360}{200}\right)}^{1/4}-1\approx 15.8\%$ per year — a figure that need not match any single year's actual growth rate, since real year-to-year growth can fluctuate even while the multi-year average (CAGR) stays fixed.

Combined two-step computations — find a missing value using a stated growth rate, then use that value to answer the actual question — are DI's most time-consuming sub-type, precisely because the two steps must both be executed correctly and in the right order. Identifying that a question needs two steps, before starting the arithmetic, avoids the wasted attempt of trying to answer it in one.

Index or base-year data expresses every value as a percentage of one chosen reference year, rather than in absolute units, and questions built on it are really percentage-change questions in disguise. If a price index reads 100 in 2018 and 132 in 2023, prices rose 32% over that period overall.

The index numbers themselves are never in rupees or any other real unit, and treating an index value as an actual price (rather than a relative figure against the base year) is a direct misread of the data.

A set built on two related tables — say, a company's unit sales in one table and its per-unit price in a second — requires cross-referencing rows by a shared key (usually the year or the category name) before any computation. The two tables are rarely printed adjacent to each other by coincidence: a question asking for total revenue needs the sales figure from one table multiplied by the price from the other, matched on the same year in both.

6. Reading chart types correctly

A truncated y-axis is the single most common visual trap in bar and line charts: if the axis starts at 80 rather than 0, a change from 82 to 88 looks dramatic on the chart even though it is a modest $\frac{88-82}{82}\times 100\approx 7.3\%$ change — always compute the actual percentage change from the labelled values, never estimate it visually from bar heights alone.

Worked Examples

Example 1 (percentage vs percentage-point — easy). A company's market share was 20% in Year 1 and 25% in Year 2. By how many percentage points, and by what percentage, did the share increase?

Percentage-point increase: $25-20=5$ points. Percentage increase: $\frac{25-20}{20}\times 100=25\%$.

Example 2 (share of total — easy). A pie chart shows a company's expenses: Salaries 40%, Rent 15%, Marketing 20%, Others 25%, with total expenses ₹80 lakh. Find the combined amount spent on Salaries and Marketing.

Combined share $=40+20=60\%$ of ₹80 lakh $=₹48$ lakh.

Example 3 (weighted average — medium). A student's marks in Physics, Chemistry and Maths are 72, 88 and 96, weighted in the ratio 1:1:2 (Maths counted double). Find the weighted average.

Total weight $=1+1+2=4$. Weighted average $=\frac{72(1)+88(1)+96(2)}{4}=\frac{72+88+192}{4}=\frac{352}{4}=88$.

Example 4 (CAGR — hard). A company's revenue grows from ₹200 crore to ₹360 crore over 4 years. Find the CAGR, and state whether this means the revenue grew by exactly the same percentage every year.

$\text{CAGR}={\left(\frac{360}{200}\right)}^{1/4}-1\approx 1.158-1=15.8\%$ per year. This is a smoothed average rate — it does not imply each individual year grew by exactly 15.8%; the actual year-to-year growth could have been, for instance, 30% one year and close to 5% another, averaging out to the same overall 4-year multiple.

Example 5 (combined two-step — hard). In a table, Region A's sales in 2022 were ₹150 crore, growing by 20% to reach the 2023 figure. Region B's 2023 sales were 90% of Region A's 2023 sales. Find Region B's 2023 sales.

Step 1: Region A's 2023 sales $=150\times 1.20=₹180$ crore. Step 2: Region B's 2023 sales $=180\times 0.90=₹162$ crore.

Example 6 (chart-reading trap — medium). A bar chart's y-axis runs from 80 to 100 (not from 0). Two bars show values 84 and 92. A student estimates visually that the second bar is "roughly double" the first based on their heights on the truncated axis. Find the actual percentage difference and explain the error.

Actual percentage difference $=\frac{92-84}{84}\times 100\approx 9.5\%$ — nowhere near "double." The visual impression of "double the height" comes entirely from the axis starting at 80 rather than 0, which exaggerates the apparent gap between two close values; the correct method is always to compute from the labelled numbers, never to estimate from bar heights on a truncated axis.

Example 7 (index numbers — medium). A price index reads 100 in 2018 and 132 in 2023. If the actual price in 2018 was ₹50, find the actual price in 2023.

The index rose 32% overall (from 100 to 132), so the actual price also rose 32%: $50\times 1.32=₹66$.

Example 8 (two-table cross-reference — hard). Table 1 gives units sold: Product X = 1,200 units in 2023. Table 2 gives per-unit price: Product X = ₹450 in 2023. Find Product X's 2023 revenue.

Revenue $=$ units $\times$ price $=1200\times 450=₹5,40,000$ — found only by matching "Product X, 2023" across both tables, not from either table alone.

Summary

DI sets share one table or chart across 4-6 questions — read every question first to know which parts of the data actually matter, then do one careful structural read of the table before answering.

Percentage change and percentage-point change are different quantities from the same data; always match the computation to the question's exact wording.

Share-of-total questions reduce to one division once a total row or column is available; weighted averages require the stated weights, not a naive row-count average.

CAGR smooths a multi-year change into one constant annual rate and does not imply uniform year-to-year growth; combined two-step questions require identifying both steps before starting the arithmetic.

Never estimate a percentage change visually from a chart — a truncated axis reliably exaggerates small differences, and the labelled values must always be used for the actual computation.