Data Interpretation Tricks for SSC, RRB and IBPS: Tables, Pie Charts, Line Graphs and Caselets

You open a DI set in the mock, see a table with five companies and three years, and your stomach drops a little. Five questions are hanging off it. Each one wants a percentage, a ratio or an average, and you have about ninety seconds per question if the section is going well. Most people either freeze for a minute reading the table or start calculating exact values they never needed.

Data interpretation is one of the few places in the quant syllabus where a plain method beats raw calculation speed. The numbers in these sets are chosen to be friendly. If you find yourself multiplying 4-digit numbers, you are usually doing it the hard way.

This post is about the working habits that make DI fast: how to read a set before touching any question, how to avoid exact division, how to handle each chart type, and which traps the paper setters like. Every example below has been calculated and checked, so you can copy the method with confidence.

Where DI shows up, and how much of your score it can carry

The place you meet DI depends on the exam, and the pattern changes now and then, so treat the numbers here as a guide and confirm against the current notification.

• SSC CGL Tier 1: there is no separate DI section. The 25-question quant section carries a few DI questions (commonly a table, a bar graph or a pie chart, often two to five questions in the section, depending on the shift). In Tier 2 the quant paper also includes data-based questions.
• RRB NTPC CBT 1: the 30-question maths section usually has a small DI cluster, typically 2 to 3 questions in recent papers according to the analyses circulated by coaching sites. Not official, but consistent with what students report.
• IBPS PO and Clerk prelims: the quant section (35 questions in a 20-minute slot for PO prelims) always carries a DI set or two, and often three sets of 5 questions. That is a big share of the section.
• IBPS PO Mains: the Data Analysis and Interpretation section is a full section (recent pattern write-ups give 35 to 40 questions in about 45 minutes, and the pattern has been revised more than once) where DI is the heart of it, along with data sufficiency, quantity comparison and some arithmetic. Check the current notification on ibps.in for exact numbers.

So if you are an SSC aspirant, DI is a modest slice you can practise cheaply. If you are a banking aspirant, it is one of the deciding areas, and mains DI in particular is a pure speed and selection game.

Read the set before the questions

The single most useful habit: spend ten or fifteen seconds on the set itself.

1. Read the title and the units. "In thousands", "in crore", "percentage of total" and "per cent profit" are different worlds. A surprising number of wrong answers come from mixing a value with a percentage.
2. Look at the row and column headings. Which years? Which categories?
3. Notice whether any figure is an anchor. Pie charts often give one absolute value ("total expenditure is Rs 7,20,000"), and everything else hangs off it.
4. Skim the questions, not to solve them but to see what the set is asking for. If three of five questions ask for a percentage change of the same row, you will work it out once.

If you begin calculating before doing this, you calculate the same thing twice, or you calculate something the question does not ask for.

The four calculations that cover most DI questions

Almost everything reduces to four operations, and each has a shortcut.

1. Percentage change

The formula is change divided by the old value. The shortcut is to think in easy fractions instead of decimals.

Fraction	Percentage	Fraction	Percentage
1/2	50%	1/8	12.5%
1/3	33.33%	1/9	11.11%
1/4	25%	1/11	9.09%
1/5	20%	1/12	8.33%
1/6	16.67%	1/16	6.25%
1/7	14.28%	1/20	5%

You do not need the whole table memorised on day one, but 1/3, 1/6, 1/7, 1/8, 1/9, 1/11 and 1/12 are worth getting into your fingers. In DI they show up over and over because the data is built from round numbers.

Worked example. Take company A's sales, in lakh units, over three years: 120, 150, 135.

• 2022 to 2023: change is +30 on a base of 120. Thirty over 120 is 1/4, so 25% increase.
• 2023 to 2024: change is −15 on a base of 150. Fifteen over 150 is 1/10, so 10% decrease.
• 2022 to 2024 overall: change is +15 on 120, which is 1/8, so 12.5% increase.

Notice that the overall change (12.5%) is not the sum of the two steps (25 − 10 = 15). If you did add them, you fell into the classic trap. A 25% rise followed by a 10% fall gives 1.25 × 0.90 = 1.125, which is a 12.5% rise. Check: 120 × 1.25 = 150, then 150 × 0.9 = 135. Yes.

2. Ratio

Do not convert to decimals. Reduce by common factors and leave it.

Company A's 2024 sales are 135 and company B's are 108. The ratio is 135 : 108. Both divide by 27: 5 : 4. That took five seconds. Converting to 1.25 and then hunting through options in decimals takes longer and invites errors.

Another: D's 2024 sales (90) and D's 2022 sales (60). Ratio is 3 : 2. If an option says 1.5, you accept it. Do it in the form the options are written in.

3. Average

Use the deviation method. Pick a convenient middle number, add the small differences, divide by the count.

The 2024 sales of the five companies are 135, 108, 216, 90 and 132. Guess an average of 130. Deviations are +5, −22, +86, −40, +2. The sum is 5 − 22 + 86 − 40 + 2 = 31. Divide by 5 gives 6.2. So the average is 130 + 6.2 = 136.2. Check by direct addition: 135 + 108 + 216 + 90 + 132 = 681, and 681 ÷ 5 = 136.2. It matches.

Directly adding five 3-digit numbers is fine too. The deviation method matters when there are eight or ten figures. Then adding all is slow and easy to fumble.

For the average of one company across years: company B sold 80, 96, 108. Total is 284, divided by 3 is 94.67 (approximate). If the options are far apart you never need the decimals. Estimate 95, done.

4. Share (percentage of total)

For share questions, the part over the total is the whole game, and this is where estimation earns its keep.

Total sales in 2024 are 681. Company C sold 216. Share is 216/681. Exact division is annoying. Notice that 681 is close to 675, and 675 = 27 × 25. And 216 = 27 × 8. So the fraction is about 8/25 = 32%. The true value is 31.7%, so if the options are 28%, 32%, 36%, 40%, you know the answer without doing any long division.

The habit here: look for a nearby number that shares a factor with the numerator. It works more often than you would expect, because setters like round-ish data.

Table DI: a full worked set

Here is a compact set to practise on. Sales of five companies, in lakh units.

Company	2022	2023	2024
A	120	150	135
B	80	96	108
C	200	180	216
D	60	75	90
E	150	165	132

Totals by year are 610, 666 and 681. Let us take questions that resemble what actually gets asked.

Q1. Which company recorded the highest percentage growth from 2022 to 2024?

Compute only the growths, and use fractions. A is +12.5%. B is +28 on 80, which is 35%. C is +16 on 200, which is 8%. D is +30 on 60, which is 50%. E is −18 on 150, which is −12%. The answer is D at 50%.

You can save time by eyeballing. Only two companies could plausibly be high: B and D. C grew from 200 to 216, obviously small. E fell. A grew by 15 on 120. B by 28 on 80 and D by 30 on 60. Compare 28/80 with 30/60: the second is clearly larger. You did not compute A, C or E fully.

Q2. In which year did the share of company A in total sales come out lowest?

Shares: 120/610 = 19.67%, 150/666 = 22.52%, 135/681 = 19.82%. The lowest is 2022, by a small margin (19.67 against 19.82). This is a question where close numbers are the trap. You cannot eyeball it. Cross-multiply instead: is 120/610 smaller than 135/681? Compare 120 × 681 with 135 × 610. That is 81,720 against 82,350. The left is smaller, so 2022's share is smaller. Cross-multiplication for closely matched fractions is safer than converting both to percentages by hand.

Q3. What is the percentage point drop in A's share from 2023 to 2024, and what is the percentage decline of that share?

Points: 22.52 − 19.82 = 2.70 percentage points. Percentage decline: 2.70 divided by 22.52 is about 12%. A's own sales fell 10%, but its share fell about 12%, because the total went up (666 to 681) while A went down. This kind of question checks whether you know the difference between "percentage" and "percentage points". If the question says "percentage points", subtract. If it says "percent", divide by the old value.

Q4. The average sales of company E over three years is what percent of company D's average?

E: (150 + 165 + 132) / 3 = 447 / 3 = 149. D: (60 + 75 + 90) / 3 = 225 / 3 = 75. So 149 / 75 = 1.9867, about 198.7%, close to 200%. If the options are 150%, 175%, 199%, 250%, take 199%. In options like these, the setter has left a close-to-double answer. Round E to 150 and D to 75 to see quickly that it is 200%, then choose the nearest.

Q5. If company C's 2025 sales are 25% more than in 2024, what will they be, and what is the ratio to A's 2024 sales?

C 2025 is 216 × 1.25 = 270. Ratio to A's 135 is 270 : 135, which is 2 : 1. Nice and clean, which is a signal you did it right. If your ratio ends up being 2.0317, you have made an error somewhere.

That last point applies broadly. In DI, correct answers tend to be clean. If you land on an ugly number, recheck the values you read from the table before you recheck the arithmetic.

Pie charts

Pie charts come in two forms: percentages on the slices, or angles on the slices. Angles are the version that costs time.

Rule: 360 degrees is 100%, so one percent is 3.6 degrees. Conversely, one degree is 1/360 of the whole. You rarely want to convert angles into percentages at all. Work with the fraction of 360.

Example. A family's monthly expenditure is Rs 7,20,000 in a hypothetical set (a big number, but it makes the fractions easier to see). The pie chart shows these angles: rent 90°, food 108°, education 72°, travel 54°, savings 36°. The angles add to 360, which is worth checking first, because sometimes the last slice is not given and you have to find it.

• Rent is 90/360 = 1/4 of the total, so Rs 1,80,000.
• Food is 108/360 = 3/10, so Rs 2,16,000.
• Education is 72/360 = 1/5, so Rs 1,44,000.
• Travel is 54/360 = 3/20, so Rs 1,08,000.
• Savings is 36/360 = 1/10, so Rs 72,000.

Those add to 1,80,000 + 2,16,000 + 1,44,000 + 1,08,000 + 72,000 = 7,20,000. Good.

Now questions.

What is the ratio of food to travel expenditure? Just use the angles: 108 : 54 = 2 : 1. You never needed the total.

This is the most valuable pie-chart trick. When the question asks for a ratio or a comparison between slices, the total cancels out, so ignore it. Only bring in the total when the question asks for an actual amount.

If savings rise by 50% and everything else stays the same, what are the new savings? 72,000 × 1.5 = 1,08,000. That equals travel spending, a fact you can spot without any working.

If the total expenditure includes rent and the family's rent is increased by Rs 18,000, what is the central angle for rent in the new pie? New rent is 1,98,000 and the new total is 7,38,000 (since rent is part of the total). Angle is 1,98,000/7,38,000 × 360 = 96.6 degrees approximately. Here is where ugliness is expected: the setter has broken the round numbers deliberately, and the options will be spaced accordingly. Rough check: 1,98,000 is about 27% of 7,38,000, and 27% of 360 is 97. So 96 to 97 is right.

A note on traps. Pie charts sometimes show two pies side by side with different totals. A slice that is the same size in both pies does not mean the same amount. Always tie a percentage to its own total.

Line graphs and bar graphs

Bar and line graphs both display a trend, and the questions revolve around change. The efficient way to read them is to write the values down as a short list in your rough sheet, in order, before answering anything. Reading the graph again for each question wastes time and creates misreads.

Example line graph: production of a factory in thousand units for 2019 to 2023 reads 40, 50, 45, 60, 75.

• Average production: 40 + 50 + 45 + 60 + 75 = 270, and 270/5 = 54 thousand.
• Year-on-year percentage changes: 2020 is +10 on 40, so +25%. 2021 is −5 on 50, so −10%. 2022 is +15 on 45, which is 1/3, so +33.33%. 2023 is +15 on 60, which is 1/4, so +25%.

Now the usual question: In which year was the percentage growth highest? The absolute rise was the same (15) in both 2022 and 2023, but the base was smaller in 2022, so the growth was larger: 33.33% against 25%. Many people pick by looking at which segment of the graph is steepest and get it wrong when the vertical distance looks equal. Steepness is absolute rise. The question asked for percentage rise, so the base matters.

By what percent did production in 2023 exceed the average of the first three years? Average of first three: (40 + 50 + 45)/3 = 45. 2023 is 75. Excess is 30 over 45, which is 2/3, so about 66.67%. Check: 45 × 1.6667 = 75. Yes.

For bar graphs with two bars per year (say, imports and exports), the questions frequently ask for the difference. Write the differences down once for all years. You may need only three of the five, but the reading pass costs the same.

Caution with scales

Y-axes sometimes do not start at zero. A bar that looks double the height of another may only be 20% larger. Read the values printed on the axis. Never estimate from bar height.

Caselets: the DI set that has no chart

Caselets are paragraphs of data. They are dreaded, but only because people try to hold everything in their head. Put it into a small table on your rough sheet. That is all the trick is.

Example. In a class of 60 students, 40% are boys. 75% of the boys and 50% of the girls pass an exam.

Boys: 40% of 60 = 24. Girls: 36. Boys who pass: 75% of 24 = 18. Girls who pass: 50% of 36 = 18. Total who pass is 36, so 36/60 = 60% of the class passed. And, interestingly, an equal number of boys and girls passed (18 each) even though there are fewer boys.

A small table:

	Boys	Girls	Total
Students	24	36	60
Passed	18	18	36
Failed	6	18	24

Once that table exists, any question is a lookup. "What percent of failures are girls?" 18 out of 24, which is 75%. "What is the ratio of boys who failed to total who failed?" 6 : 24 = 1 : 4.

Start by writing the totals and category splits. When information is given in the form "X% of A and Y% of B", convert to actual counts with a convenient total. If no total is given, assume 100 or a common multiple of the denominators. It will not change ratios or percentages.

Profit and cost DI (a variant that trips people)

Some sets give income and profit percent, and ask for expenditure. Here the rule is simple, but the direction of the percentage is what confuses.

Example. A company's income in a year is Rs 480 crore and its profit is 25% (on expenditure). Expenditure is income divided by 1.25, which is 384 crore. Profit is 480 − 384 = 96 crore. Check: 96/384 = 0.25. It matches.

The trap is calculating 25% of 480 = 120 as the profit. That would be right if profit were expressed on income. Profit percentage is on cost unless the question says otherwise. If you have not yet gone through profit and loss thoroughly, the companion post on profit and loss tricks covers these direction problems in depth.

Speed: how banking DI actually gets done in the time

For the IBPS PO prelims, you get a 20-minute quant section for 35 questions. Assume a DI set of 5 questions takes about 5 to 6 minutes if you are fluent. Skip anything that needs a second pass.

A working order that many people settle on:

1. Scan the section for the DI sets first. Decide which set looks lightest (simple table, or bar graph with easy numbers).
2. Solve the light DI set first, since it carries five questions for one reading effort.
3. Then solve short arithmetic and simplification.
4. Leave the heavy caselet and any data sufficiency for the end.

This is a personal call and you should test it on mocks. The point is that sets are worth more per minute than stray questions because the reading cost is shared. A heavy set with tough numbers is the opposite: it costs a lot and you may still miss two of five. In the mains section, where every question is a data question, the set-by-set choice matters even more.

Attempt count differs, and the safe number depends on your accuracy. Negative marking in these exams is a quarter mark per wrong answer in the banking prelims, and half a mark in SSC Tier 1. Please check the current notifications. That difference changes how much you can gamble on an estimation-based answer.

Approximation inside DI

DI answers are rarely exact. Options are typically at least 3 to 5% apart. So it is fine to round.

Techniques:

• Round to the nearest factor-friendly number. Not the nearest ten, the nearest number that divides cleanly. 681 to 675 was the example above.
• Round in opposite directions when dividing. If you round the numerator up and the denominator down, the error compounds. Round them the same way, or round so the errors offset.
• Use the ends of the range. If the exact value is between 31% and 33%, and the options are 28, 32, 36 and 40, you are done. You did not need a decimal place.
• When two options are close, compute; when they are far, estimate. Look at the options first.

There is a full treatment of this in the simplification and approximation tricks post, which is worth reading alongside this one because DI is basically approximation with a chart in front of it.

The traps setters like

I have seen these in almost every set of PYQs and mocks.

• Percentage versus percentage points, as in question 3 above.
• Base confusion: "A is what percent more than B" uses B as the base. "B is what percent less than A" uses A as the base. The two answers differ. If A is 125 and B is 100, A is 25% more than B, but B is 20% less than A.
• Units: a table in thousands and a question in lakh, or a pie showing percentages while the question asks for a count.
• Incomplete pie: a slice given as "others" or missing. Total must be 360 or 100.
• "At least" or "at most" wording: questions asking whether a statement is true for all years need you to check all years. One counterexample makes the statement false.
• Average of averages: the average of two group averages is the overall average only if the groups are equal in size.
• Data that is not given. Some options say "cannot be determined." These are correct more often than students expect, when the graph shows only a percentage and the question asks for an absolute count without a base.

Practising DI properly

Doing dozens of DI sets without a review does little. A better routine, if you have, say, six weeks:

• Week 1-2: Build the fraction-percentage table into memory and do 20 tables with untimed working, but write only the necessary calculations.
• Week 3-4: Timed sets. Five questions in seven minutes at first. Then six. Then five. Track which questions you skipped or got wrong and why (misread, wrong formula, arithmetic).
• Week 5-6: Mixed sets: caselets, pie plus table, and two graphs in one set. Also do a few data sufficiency questions if you are aiming for mains.

Most of the improvement comes from the review step. When you get one wrong, decide whether you made a reading error, a method error or a speed error. Reading errors are the most common, and they do not go away by learning more formulas.

Pareeksha's sectional practice has quant sets you can time yourself on, and after each full mock in /mock-tests the percentile and section breakdown shows you whether DI is where your marks are leaking. If you are attempting only the easy set and skipping the rest, the breakdown will tell you so.

A last look at a realistic exam-style set

Try this one yourself before reading the answer. A pie chart shows the distribution of 1,200 candidates who took a test, by state: State P has an angle of 120°, State Q 90°, State R 60°, State S 45° and State T 45°. The pass percentage is: P 60%, Q 80%, R 50%, S 40%, T 80%.

Angles add to 360, good. Candidates: P = 1/3 of 1,200 = 400. Q = 1/4 = 300. R = 1/6 = 200. S = 1/8 = 150. T = 1/8 = 150. Those total 1,200.

Passed: P 60% of 400 = 240. Q 80% of 300 = 240. R 50% of 200 = 100. S 40% of 150 = 60. T 80% of 150 = 120. Total passed = 240 + 240 + 100 + 60 + 120 = 760.

Overall pass percentage = 760/1,200 = 63.33%. Since 1,200 × 0.6333 = 760, that is confirmed.

Questions you can pose yourself: which state contributed the most candidates who passed? P and Q tie at 240. What is the ratio of those who failed in R to those who failed in S? R failed 100, S failed 90. Ratio is 10 : 9. Failed in P is 160. In T it is 30. Which state has the fewest failures? T, with 30.

If you got those without opening a calculator, you have the method.

What to do this week

Take one table-based set from a mock you have already given and redo it slowly, writing down only the percentages and ratios you actually needed. Then time yourself on a fresh set. If you are also preparing for banking mains, add the caselet routine above.

Once you are comfortable here, the natural next posts are percentage tricks and maths chapter weightage to decide how much of your quant time should go to DI at all.

FAQ

Is DI easier than the rest of quant? It is easier in the sense that no new formula is needed. It is harder in the sense that it is long. Time is the constraint, not concept.

Should I use the on-screen calculator in bank exams? IBPS exams have provided an on-screen calculator in some sections, but do not depend on it. The notification and instructions on the day tell you. Practise without one because it is slower than estimating anyway.

How many DI sets should I practise per day? Two to three sets, done properly and reviewed, are worth more than ten skimmed. Speed comes over weeks.

What if I misread the data in the exam? It happens to everyone. Build a habit of quickly tapping back to the table before answering close-call questions. It costs five seconds.

Do I need to memorise squares and fractions? Squares up to 30 and the fraction-percentage table above are worth it. Beyond that, use estimation.

Are the numbers in your examples from real exams? No. They are made-up sets for practice, built so that every step can be checked. The question styles resemble what the exams ask, but for actual papers see the previous year papers on the official sites.