Analytics & Data Science Interview Question Bank by Chicago Booth Analytics Club
 Share
The version of the browser you are using is no longer supported. Please upgrade to a supported browser.Dismiss

 
View only
 
 
ABCDEFGHI
1
VCompany /
Organization
Role
Interview
Round
Interviewer(s)
Interview
Medium
Question
Type
Question
2
A/B Testing
3
Facebook
Data Scientist
(University Grad)
Round 1
Data Scientist /
Data Engineer
Phone
4
Facebook
Data Scientist
(University Grad)
Round 1
Data Scientist /
Data Engineer
PhoneProduct
Sense
Propose metrics to measure the health of the Messenger app
5
Facebook
Data Scientist
(University Grad)
Round 1
Data Scientist /
Data Engineer
PhoneProduct
Sense
If we detect that the volume of messages have fallen 5% in the past 2 days,
how do you diagnose the root cause of the decline?
6
Facebook
Data Scientist
(University Grad)
Round 1
Data Scientist /
Data Engineer
PhoneA/B Testing
Assume we're thinking about making a certain icon to serve as the Send button
in the Messenger app, how do you know if the change is gonna be beneficial?
7
Facebook
Data Scientist
(University Grad)
Round 1
Data Scientist /
Data Engineer
PhoneSQL
From a SQL table with user_id, content_id, target_content_id,
generate a table for the distribution of the number of comments per user.

(target_content_id is a foreign key for content_id; if target_content_id is not NULL,
then the concerned content is a comment attached to the target main content)

If there is an additional column user_type classifying the users into a number
of categories, how do you modify the SQL statement to generate the distribution
of the number of comments for each user type?
8
Facebook
Data Scientist
(University Grad)
Round 1
Data Scientist /
Data Engineer
PhoneProduct
Sense
How do we detect post + comment threads that are of congratulatory nature?
If we use the technique you propose, describe the likely false positive cases.
9
10
LyftAnalystRound 2
11
LyftAnalystRound 2Product
Sense
In a dassboard that you are presenting to CEO, what 5 metrics will you use to showcase
the health and growth of Lyft?
12
LyftAnalystRound 2EconomicsHow do you calculate customer Lifetime Value (LTV)?
13
LyftAnalystRound 2SQL
14
LyftAnalystRound 2
Tell me about an analysis you did, how you approached it,
presented it and how you achieved buy-in.
15
LyftAnalystRound 3
16
LyftAnalystRound 3Walk through an impactful project you've worked on.
17
LyftAnalystRound 3SQL
Given 2 tables, A with id and rev; B with id and Rev,
create output that has all of the id and rev of B and the additional ones from A that weren't in B
18
LyftAnalystRound 3EconomicsHow would you price a referral program?
19
20
MyFitnessPalProduct AnalystRound 1
21
MyFitnessPalProduct AnalystRound 1Product
Sense
If a PM came asking to compare calories/meal between US/UK - what would you do?
22
MyFitnessPalProduct AnalystRound 1Statistics /
Econometrics
If the distribution of a variable is non-normal, what would you do?
23
MyFitnessPalProduct AnalystRound 1Statistics /
Econometrics
How do you test for significance of difference between data sets?
24
MyFitnessPalProduct AnalystRound 1Statistics /
Econometrics
There are 2 cities, 1 with large hospital, one with small hospital.
A hospital has ratio of staff is 60% Male to 40% Female; which hospital is this more likely to be?
25
26
Nest LabsRetail AnalystFinal
27
Nest LabsRetail AnalystFinalCase
Sales has negotiated display space in 200 new (out of 1000) stores,
how do you choose which 200 stores to go to?
28
Nest LabsRetail AnalystFinalSQL
2 tables. T1: user_id, name, refered_by name.
T2 Purchase $ and user_id.
Get $, name and referred by name.
29
Nest LabsRetail AnalystFinalTake-HomeWhy would sales increase and store level efficiency drop?
30
Nest LabsRetail AnalystFinalTake-Home
1 store has profit at $1000/week and other at $100/week,
merchandising (physical fixtures) cost $1000, justify ROI
31
Nest LabsRetail AnalystFinalTake-HomeDetermine store-level weeks of inventory?
32
Nest LabsRetail AnalystFinalBehavioral
Example of prior-experience that fits into the job description I described - tactical, cross-functional analytics
33
Nest LabsRetail AnalystFinalBehavioralAnalytical Insight that you are most proud of.
34
Nest LabsRetail AnalystFinalCase
Launching a new product never done before, forecast sales. What info do you need, where would you get it?
35
Nest LabsRetail AnalystFinalBehavioralHow have you managed team with opposing incentives?
36
Nest LabsRetail AnalystFinalBehavioralWhy Nest, Why this role?
37
Nest LabsRetail AnalystFinalBehavioralExample of relavent experience from past job?
38
Nest LabsRetail AnalystFinalBehavioralHow would you get grounded here at Nest in 30 days?
39
Nest LabsRetail AnalystFinalCase
How do you balance the trade off between Profit and Rev,
specifically - if you had a chance to participate in Home Depot promo which was profit-negative,
what would you do?
40
Nest LabsRetail AnalystFinalBehavioralTell me a time you worked with a challenging co-worker.
41
Nest LabsRetail AnalystFinalBehavioralWhat values do you look for in an org?
42
Nest LabsRetail AnalystFinalBehavioralWhat does your inbox look like?
43
Nest LabsRetail AnalystFinalCase
You have all these sources of data (internal, external, syndicated)
and are asked to manage BestBuy's account, How would you prioritize your first month?
44
Nest LabsRetail AnalystFinalCaseTell me about the most complex problem you've worked on.
45
Nest LabsRetail AnalystFinalBehavioralGive me your proudest leadership example.
46
Nest LabsRetail AnalystFinalCaseWhat would your co-workers say are your strengths and weaknesses?
47
Nest LabsRetail AnalystFinalBrainteaser
A, B and C go camping.
A buys 5 logs of wood, B buys 3 logs of wood. C gives them both $8
how should they split up the $8 assuming all shared heat equally?
48
49
Omada HealthAnalystRound 2
50
Omada HealthAnalystRound 2Product
Sense
In a program with 3-month signup winodw, more recent data will have less weight loss.
What would you do to keep the customer?
51
Omada HealthAnalystRound 2Product
Sense
Average in our data set is 5%, you do new pilot and avg is 2%, diagnose the issue.
52
53
Omada HealthAnalystFinal
54
Omada HealthAnalystFinalBehavioralTime you worked with a difficult person
55
Omada HealthAnalystFinalBehavioralTime you worked with clients
56
Omada HealthAnalystFinalBehavioralTime you worked with internal stakeholders
57
Omada HealthAnalystFinalBehavioralHow do you go about figuring how to craft that story
58
Omada HealthAnalystFinalBehavioralExample of work/story that you think would be relevant
59
Omada HealthAnalystFinalBehavioralExample when the data was not favorable, how you presented that to client
60
Omada HealthAnalystFinalBehavioralHow would your coworkers describe you
61
Omada HealthAnalystFinalBehavioralHow do you receive feedback?
62
63
Uber
Data Scientist
(Economics / Econometrics),
Dynamic Pricing
Round 2-3, after
Take-Home Exercise
Lead Data Scientist,
Dynamic Pricing
Phone
64
Uber
Data Scientist
(Economics / Econometrics),
Dynamic Pricing
Round 2-3, after
Take-Home Exercise
Lead Data Scientist,
Dynamic Pricing
PhoneProduct
Sense
Why is it important for Uber to have interpretable regression models predicting Demand and Supply
using Price and other relevant variables? Why can't we just build huge, complex Machine Learning
models that predict Demand and Supply volumes well according to a simple test-set RMSE criterion?

Why must we spend so much time modeling carefully to obtain sensible numerical estimates
for the coefficents on Price?
65
Uber
Data Scientist
(Economics / Econometrics),
Dynamic Pricing
Round 2-3, after
Take-Home Exercise
Lead Data Scientist,
Dynamic Pricing
PhoneStatistics /
Econometrics
What endogeneity exists in a model that regresses the Demand conversion rate (i.e. how many Riders
accept the quoted fare estimate) on Price and Waiting Time?

What happens when you have such endogeneity? How can you fix this endogeneity problem?
66
Uber
Data Scientist
(Economics / Econometrics),
Dynamic Pricing
Round 2-3, after
Take-Home Exercise
Lead Data Scientist,
Dynamic Pricing
PhoneStatistics /
Econometrics
How do you build a Demand prediction model that responds / adjusts quickly to unexpected
surges in demand (e.g. from a sudden rain storms)?

If the approach is complex, how can you explain it clearly to a Product Manager?
67
Uber
Data Scientist
(Economics / Econometrics),
Dynamic Pricing
Round 2-3, after
Take-Home Exercise
Lead Data Scientist,
Dynamic Pricing
PhoneStatistics /
Econometrics
Discuss how would you factor in the effects of expectations in modelling Demand and Supply?
68
Uber
Data Scientist
(Economics / Econometrics),
Dynamic Pricing
Round 2-3, after
Take-Home Exercise
Lead Data Scientist,
Dynamic Pricing
PhoneEconomics
Suppose Taylor Swift is in town and hoards of people are going to see her.
Assume for simplicity that, exogenously, some folks at Uber decide that at the very moment
Swift's concert ends, the price level of Uber rides in that local area should be 3x the normal price.

Describe the dynamics of the market price in that area leading to, at, and after that moment.
Which market agents are facing which trade-offs? How do those trade-offs result in the dynamics
you've just described?
69
70
Uber
Data Scientist,
Intelligent Decision Systems
/ Anomaly Detection
Round 1, before
Take-Home Exercise
Lead Data Scientist,
Intelligent Decision Systems
/ Anomaly Detection
Phone
71
Uber
Data Scientist,
Intelligent Decision Systems
/ Anomaly Detection
Round 1, before
Take-Home Exercise
Lead Data Scientist,
Intelligent Decision Systems
/ Anomaly Detection
PhoneStatistics /
Econometrics
How can we detect anomalies in time series?
72
Uber
Data Scientist,
Intelligent Decision Systems
/ Anomaly Detection
Round 1, before
Take-Home Exercise
Lead Data Scientist,
Intelligent Decision Systems
/ Anomaly Detection
PhoneStatistics /
Econometrics
If we use a huge number of variables in a predictive model, how can we know / have a good guess
at which variables are relevant and which are not?
73
Uber
Data Scientist,
Intelligent Decision Systems
/ Anomaly Detection
Round 1, before
Take-Home Exercise
Lead Data Scientist,
Intelligent Decision Systems
/ Anomaly Detection
PhoneStatistics /
Econometrics
What variables do we need to take into account to have a good model estimating
the probability of a Driver accepting a ride request?
74
75
PremiseSales EngineerRound 2
76
PremiseSales EngineerRound 2Statistics /
Econometrics
How would you spot an outlier time series?
77
PremiseSales EngineerRound 2Statistics /
Econometrics
Calculate expected value of ratio boys/girls in a country that is sexist
and birth kids till you have a boy then stop.
78
PremiseSales EngineerRound 2Y = mx versus x=my, how do graphs differ?
79
80
PremiseSales EngineerFinal
81
PremiseSales EngineerFinalBehavioralExample of time working with a client unsavvy with technology.
82
PremiseSales EngineerFinalAlgorithm /
Optimization
Given array of values and freq, write a funct that generates the values with given probs.
83
84
Yelp
Product Manager
(Analytics)
Screener
85
YelpProduct Manager
(Analytics)
ScreenerSQLFind the # of rows that match business_id = 1000
86
YelpProduct Manager
(Analytics)
ScreenerSQLWhat is a database index?
What are issues with having too many indices?
87
88
Yelp
Product Manager
(Analytics)
Round 1
89
YelpProduct Manager
(Analytics)
Round 1Why Yelp; What did you do in XYZ PM role before?
90
YelpProduct Manager
(Analytics)
Round 1Software /
Technology
What data analysis stack do you use (excel/tableau) / Python / R;
Why do you choose this?
91
YelpProduct Manager
(Analytics)
Round 1Software /
Technology
How would you collaborate on analysis? What tools and workflow?
92
YelpProduct Manager
(Analytics)
Round 1Product
Sense
If you were given 1 week to play with all of Yelp's data, what would you do?
93
YelpProduct Manager
(Analytics)
Round 1Product
Sense
Jeremy asked you to find top 10 brunch places in USA, how would you answer?
94
95
zzz ANON. CO.Data ScientistRound 1-2Software Engineer
Google
Hangout
96
zzz ANON. CO.Data ScientistRound 1-2Software EngineerGoogle
Hangout
Describe a typical Data Science project's phases and workflows
97
zzz ANON. CO.Data ScientistRound 1-2Software EngineerGoogle
Hangout
Machine
Learning
Explain how Principal Components Analysis (PCA) works
98
zzz ANON. CO.Data ScientistRound 1-2Software EngineerGoogle
Hangout
Numerical
Estimation
Use as simple tools as possible, give an estimate for the distance between the Earth and the Moon
99
zzz ANON. CO.Data ScientistRound 1-2Software EngineerGoogle
Hangout
Algorithm /
Optimization
If we have a random number generator G that generates integers 1 to N uniformly,
discuss an algorithm to create a random number generator F that generates integers 1 to M
uniformly
100
Loading...
 
 
 
INTERVIEW QUESTION BANK
Companies / Organizations
Question Types
Models/Methods Catalog