Case Questions for Data Science

Interviews involving case studies are quite common for many job roles. For data science roles, these have evolved to mimic the work the data scientist will be doing if hired. They’re a great way for companies to simulate a data science project and assess how a candidate will perform in the real world.?

Before we dive into any practice questions, let’s understand what interviewers are looking for in a case interview and how to best tackle them.?

What is an interviewer testing you on in a case interview?

• Ability to synthesize information—Interviewers are looking to see if you can piece together information on multiple fronts and come up with coherent and concise ideas. Most case interviews begin with a broader prompt giving an overarching view of the problem statement. The interviewer is looking at whether you can break down this top-level information and work toward a solution by putting together the jigsaw pieces along the way.?
• Analytical thinking—Given the nature of the job role for a data scientist, the interviewer will be judging you on your ability to think from an analytical lens and highlight how data can be best used to solve the business problem posed to you.?
• Communication skills—Interviewers are looking for how well you communicate your approach and the findings and how you summarize the overall problem. They will also be looking at your ability to communicate with both a technical and nontechnical audience.?
• Product Sense—If you’re interviewing for a product company, they will be evaluating you on your “product sense”—what makes a product great and what doesn’t. You can practice developing your product sense by looking at everyday things critically. You may ask yourself questions like: Why is my iPhone so easy to use? What is so great about the Instagram story feature? What makes it easy to search for bed and breakfasts on Airbnb? If you would like to take it a bit further and wear the product manager hat for a while, then start to think about what can make this product better. While you may not be asked product enhancement questions in a data science interview, practicing along these lines will allow you to think outside the box and will push your creative mindset.?

What is a good approach to tackle case questions?

• Ask and understand the “What”—When you are given the case prompt, take a minute or two to carefully understand the question and the objective behind it, and confirm your understanding of what is being asked. You may not know the answer to the question off the bat, and that’s okay. Rather, that’s how it should be. You will get to the finer aspects later on, but as the first step, understand what the problem statement is and what is it that you are tasked to solve for.?
• Dive into the “Why & How”—Once you have a handle on the goals, move into the “Why & How.” For example, if it’s a profitability case, identify the drivers of profitability and then move to understanding how changes in these drivers impact profitability. Write down your structure as you do this so the interviewer can give you early feedback if you are steering away from where they want you to go.?
• Hypothesize and brainstorm options—Based on the overall goal, build hypotheses and talk through the various options as you see fit for the case. This is also a great time in the interview to ask if additional data points are available to dig in further and to validate your assumptions with your interviewer along the way.?
• Actively communicate thought process—Having given multiple case interviews as part of the recruiting process after his MBA, Ganapati Raghunathan offered the following insight: “It is key to keep communicating your thought process with the interviewer. The interviewer will usually pull the candidate back on track if they go on a tangent.”
• Summarize and recommend—Once you have evaluated different options and essentially walked the interviewer through your “decision tree,” summarize your recommendation/conclusion, list out any key assumptions you have made while arriving at the conclusion, and state any risks that this recommendation may bring in.?

Other things to watch out for:

• If in the process you ever stuck and things seem daunting, think of yourself as the interviewer’s colleague. Think of the question as if you were leading the project. This will put you right back in the driver’s seat and will give you the confidence to ask the right questions and formulate an approach.?
• If you find yourself rambling and giving long-winded answers, pause. Take a step back and ask for a couple minutes to work through your thoughts. Structure your response and write down key points and then start to answer again. Remember, there is never a wrong time to ask for a couple minutes to structure your response.
• Do not be fixated on only one area, especially if the interviewer is not keen on going in that direction. Read the cues from your interviewer as they often will nudge you in the correct direction.?
• Do not get swayed into building a perfect data-driven solution only. Validate it with your business intuition and assess whether it fulfills the broader business objective before making a recommendation.??

Stay tuned for the post next week with a mock case interview with answers!