Mu Sigma Coding and Analytical Round Questions
Last updated 2026-11-19
Entry-level analytics rounds usually test whether you can compute carefully and reason aloud rather than whether you can write advanced algorithms. This page gives twelve practice problems in that style: Python warm-ups, SQL and pandas on small tables, two guesstimates and break-even case logic, with worked solutions, complexity where it matters and the mistakes that cost marks.
They are practice problems in the style of such rounds, not copies of any test. Use them with Mu Sigma interview questions, how to crack the SQL interview and how to crack the data analyst interview.
| Languages here | Python, SQL, pandas |
|---|---|
| Also covered | Guesstimates, break-even case, puzzles |
| Level | Easy to medium |
How to approach these problems
Write the formula before any number, keep units visible, and sanity-check the answer against common sense (a market of 5 crore bottles a day in one city is a sign of a slip). In code, print intermediate results for small inputs and handle empty input explicitly.
Hiring status
We hold no sourced facts about Mu Sigma tests, drives, openings or cutoffs, so none are stated. Use the company's own careers site for current information.
Mu Sigma Coding and Analytical Round Questions: questions with worked answers
1 Python: check whether a number is prime and list the primes up to n.
easyFor a single check, test divisors up to the square root: O(sqrt n). For all primes up to n, the Sieve of Eratosthenes marks multiples of each prime: O(n log log n) time, O(n) space. Mistake: testing every divisor up to n, and forgetting that 0 and 1 are not prime.
pythondef primes_upto(n): if n < 2: return [] sieve = [True] * (n + 1) sieve[0] = sieve[1] = False for i in range(2, int(n ** 0.5) + 1): if sieve[i]: for j in range(i * i, n + 1, i): sieve[j] = False return [i for i, p in enumerate(sieve) if p] print(primes_upto(30)) # [2, 3, 5, 7, 11, 13, 17, 19, 23, 29]2 Python: the n-th Fibonacci number efficiently.
easyIterate keeping the last two values: O(n) time, O(1) space. Plain recursion is O(2^n) because it recomputes subproblems; memoisation fixes that. Mistake: starting the sequence at the wrong index; state whether F(0) is 0 or you start from 1.
pythondef fib(n): a, b = 0, 1 for _ in range(n): a, b = b, a + b return a print([fib(i) for i in range(8)]) # [0, 1, 1, 2, 3, 5, 8, 13]3 Python: check whether an integer is a palindrome without converting it to a string.
easyReverse the digits arithmetically with divmod and compare with the original; negatives are not palindromes. O(log n) time. Mistake: not copying the number before destroying it in the loop.
pythondef is_palindrome(n): if n < 0: return False rev, m = 0, n while m: m, d = divmod(m, 10) rev = rev * 10 + d return rev == n print(is_palindrome(12321), is_palindrome(123)) # True False4 Puzzle in code: 100 doors, all closed. In pass k you toggle every k-th door. Which doors are open after 100 passes?
mediumDoor n is toggled once for each divisor of n. Divisors come in pairs, except when n is a perfect square (the square root pairs with itself), so only perfect squares have an odd number of divisors and end open: 1, 4, 9, 16, 25, 36, 49, 64, 81 and 100, that is 10 doors. Simulating confirms it in O(n log n).
pythondoors = [False] * 101 for k in range(1, 101): for d in range(k, 101, k): doors[d] = not doors[d] print([i for i in range(1, 101) if doors[i]]) # perfect squares up to 1005 SQL: the average salary per department, listing only departments above 50,000, and the number of employees in each.
mediumGroup by department, compute AVG and COUNT, and filter the aggregate with HAVING. Mistake: using WHERE for an aggregate condition, which is a syntax error, and counting with COUNT(salary) when NULL salaries should still count as employees.
sqlSELECT department, AVG(salary) AS avg_salary, COUNT(*) AS employees FROM employees GROUP BY department HAVING AVG(salary) > 50000 ORDER BY avg_salary DESC;6 SQL: employees who earn more than the average salary of their own department.
mediumCompare each employee with the average of their department, either by a correlated subquery or a window function that computes the department average on every row. Mistake: comparing with the company-wide average, which answers a different question.
sqlSELECT name, department, salary FROM ( SELECT e.*, AVG(salary) OVER (PARTITION BY department) AS dept_avg FROM employees e ) t WHERE salary > dept_avg;7 pandas: for each product category compute the mean price, then keep categories whose mean exceeds the overall mean.
mediumGroup by category and take the mean, compute the overall mean from the original column, then filter. Mistake: taking the mean of the category means as the overall mean, which differs when categories have different sizes.
pythonimport pandas as pd df = pd.DataFrame({"cat": ["a", "a", "b", "c", "c", "c"], "price": [10, 20, 40, 5, 5, 8]}) overall = df["price"].mean() by_cat = df.groupby("cat")["price"].mean() print(by_cat[by_cat > overall])8 Break-even case: fixed costs are Rs 10 lakh, the price per unit is Rs 500 and the variable cost is Rs 300. How many units to break even, and what is the profit at 8,000 units?
mediumContribution per unit is 500 - 300 = Rs 200. Break-even units are fixed cost divided by contribution: 10,00,000 / 200 = 5,000 units. At 8,000 units the contribution is 8,000 x 200 = Rs 16 lakh; subtract fixed costs of Rs 10 lakh for a profit of Rs 6 lakh. Follow-ups: how does break-even move if the price falls to Rs 450? Contribution becomes 150, so break-even rises to about 6,667 units.
pythonfixed, price, var = 1_000_000, 500, 300 contribution = price - var print(fixed / contribution) # 5000.0 print(8000 * contribution - fixed) # 6000009 Quant: a shirt marked Rs 1,000 is sold at a 20 percent discount and still makes a 25 percent profit on cost. What is the cost price?
mediumThe selling price is 1,000 x 0.8 = Rs 800. Profit is 25 percent of cost, so selling price is 1.25 times cost and cost = 800 / 1.25 = Rs 640. Check: 640 x 1.25 = 800. Mistake: taking the profit percentage on the selling price instead of the cost, which gives 800 x 0.75 = 600 and a wrong answer.
10 Guesstimate: the daily market for packaged drinking water bottles in a city of 5 million people.
mediumAssume 5 million people, and that about 30 percent buy packaged water on a given day (the rest use home filters or taps): 1.5 million buyers. Assume about 1.5 bottles each, mixing 1-litre and small bottles: about 2.25 million bottles a day. At an average of Rs 15 a bottle, that is about Rs 3.4 crore per day. Sanity check: per person about 0.45 bottles a day, plausible for an urban population with mixed habits. Name the assumption to validate first: the share who buy daily.
11 Python: given a list of daily orders, find the longest streak of days where orders increased compared with the previous day, and the day it ended.
hardScan once with a current streak and a best streak; extend when today is greater than yesterday and otherwise reset to 1. Record the end index when the best improves. O(n) time, O(1) space. Mistake: updating the best only at the end of the loop, which misses a streak that runs to the last day.
pythondef best_streak(a): best = cur = 1 end = 0 for i in range(1, len(a)): cur = cur + 1 if a[i] > a[i - 1] else 1 if cur > best: best, end = cur, i return best, end print(best_streak([1, 2, 3, 1, 2, 3, 4, 2])) # (4, 6)12 Puzzle: the average of five numbers is 20. When one is removed the average of the rest is 18. What was the number removed?
hardThe total of five numbers is 5 x 20 = 100. After removing one, the four remaining total 4 x 18 = 72. The removed number is 100 - 72 = 28. Generalise: removed value = n x old average - (n - 1) x new average. Check by recomputing: (72 + 28) / 5 = 20.
Frequently asked questions
What coding is needed for Mu Sigma style rounds?
Reported preparation for analytics fresher roles stresses aptitude, puzzles and light SQL or Python. Check your role description, since technical depth varies.
How should I practise guesstimates?
Pick everyday quantities, split them into factors, state assumptions and cross-check by a second route. Write them down so you can compare later attempts.
Should I learn R?
If the job description lists R, learn the basics of data frames and summary statistics. Python with pandas covers the same ground for most candidates.
Are these actual Mu Sigma questions?
No, they are practice problems in the style of such rounds, not copies of any test.
How do I avoid arithmetic mistakes?
Use round numbers, write units, and re-derive the answer a second way before moving on.
Related guides
More in Company Interview Questions
- Mu Sigma Interview Questions with Answers
- Myntra Coding Round Questions with Solutions
- Myntra Interview Questions and Answers
- Nagarro coding round questions
- Nagarro interview questions and answers
- Netflix Coding Round Questions with Solutions in Python and Java
- Netflix Interview Questions and Answers for Engineers
- NTT DATA coding round questions
- NTT DATA interview questions and answers
- Nutanix Coding Round Questions With Solutions
- Nutanix Interview Questions and Answers
- NVIDIA Coding Round Questions with Solutions