Have you ever heard someone say, "What's your best guess?"? It certainly comes as no surprise that some guesses are better than others. Whether wrong or good, guesses provide us with at least a modicum of information as we try to decide on a course of action or a choice. Using a guess as guidance in decision-making isn't ideal, but it's better than nothing.
On the other hand, if a guess is sufficiently awful, perhaps it would have been better not to guess at all! After all, some of us are better at guessing than others.
Companies looking for a data scientist want to have people who can make good guesses since sometimes the role requires guesswork. Therefore, prospective data scientists going through the interview process need to show that they have a solid grip on making a guess, an estimate, or, putting those two terms together, a guesstimate.
That’s why today we’re covering the top guesstimate questions for data science interviews (and applicants to positions in similar, related fields!). If you want to become a data scientist, business analyst, data architect, or consultant, you need to become familiar with these kinds of questions.
But first, what in the world is a guesstimate?
What Is a Guesstimate?
The term “guesstimate” is a portmanteau of “guess” and “estimate.” Guesstimates are approximations based on the available limited information. A guesstimate is an information-based guess, not an accurate answer.
Guesstimates have the following five characteristics:
- Meaning: Understanding the problem, figuring out its purpose, and why you want to solve it
- Definition: The explanation, the object in question, and the input/output of the process flow
- Guessing: Thinking and reaching a conclusion
- Estimate: Coming up with an estimate based on the numbers you must work with
- Generate an idea: Taking the concept and implement it with research and development
Tips for Data Science Guesstimate Questions
Although guesstimate questions don’t produce pinpoint accurate answers, you can still embrace certain habits and tactics that can improve the quality of your answers. Keep these tips in mind:
- Remember, there's no such thing as a correct answer: So please don't get yourself worked up looking for a precise solution; it's not going to happen.
- Write things down: Write every part of the question down, and if the question requires segmentation, create a flowchart showing each segment. The interviewer may want to see your calculation sheet, so don’t make the sheet illegible or filled with rough calculations.
- Practice rounding: Don’t be worried about fractions or decimals—round your figures to the nearest whole number.
- Facts top feelings. Avoid relying on your gut feelings. Logic and facts (even if you have only a few) carry more weight than how you feel or believe.
- Stay unflappable: You may get a weird question, but don’t let that rattle you. Every question has an answer, no matter how strange.
- Take a moment and think things through: You don’t get points for speed. Pause and reflect on the question; quiet your mind and think things through rationally.
- Clarify your thoughts, then voice them: Once you’ve had the chance to consider all the angles and employ whatever facts at your disposal, come up with an answer in your head, then express it.
- Remember, there’s no wrong answer: This concept is important enough to repeat. There are no correct answers! If you hold off on answering until you come up with the perfect answer, you’ll be dead in the water, and the interviewer won’t be impressed.
Data Science Guesstimate Interview Questions and Answers
Here’s a collection of ten of the most common guesstimate questions for data science interviews, covering all areas of expertise, from beginners to seasoned pros. Some of the questions may cover concepts already covered, but they’re included anyway for the sake of creating a definitive guesstimate question list.
1. What is a guesstimate?
A guesstimate is a guess based on existing information; an approximation based on available data.
2. How do you solve a guesstimate question?
You can solve a guesstimate question by breaking it down into four steps:
First, clarify any unclear terms in the query.
Second, break down large numbers into smaller, easier-to-handle pieces.
Third, use background knowledge to estimate each piece.
Finally, consolidate all the parts and present your conclusions.
3. Give some examples of guesstimate questions.
Here are some typical guesstimate questions:
What is the current number of Android phones in use in Delhi?
How many square inches of pizza do Americans consume in a day?
How much tea do people in the United Kingdom drink on any given day?
How many golf balls can you fit in a MINI Cooper?
4. What function do guesstimate questions serve in Data Science?
First, they help gauge a data analyst’s capacity to understand the situation.
Second, they show the scope of the data analyst’s ability to connect things and arrive at an answer.
Third, they measure how well the analyst can prioritize or dismiss different parameters.
Finally, they show how well the analyst can work with limited data.
5. What are the two methods of approaching guesstimate questions?
Top-down method. Start with the largest universe possible (which the guesstimate is a part of), apply sets of conditions and filters, reducing the universe’s numbers into something that works for the estimate.
Bottom-up method. Start with a low-level statistic and build your way up to an answer. For instance, if you wanted to figure out a salesperson’s monthly income, you’d figure out their weekly earnings then multiply the result by 4.
6. Explain the three different guesstimates based on how you approach a solution.
The three types are:
The Household Approach: This approach tackles household-based guesstimates.
The Population Approach: This approach handles population-related guesstimate questions, such as finding out how many people live in a given area.
The Structural Approach: This approach creates guesstimates for situations like finding out how many vehicles use a particular bridge every day.
7. How would you determine how many iPhone users exist in the United Kingdom?
First, clarify that the question includes all models of iPhone. Second, determine the population. The UK has approximately 67 million people, with about 40 percent being children and the elderly, so we won’t count them. That means we’re left with 40,200,000 potential iPhone users.
Since iPhones tend to be expensive, we’ll further eliminate anyone below the middle class. About 25 percent of the UK’s population falls into the middle-class category, and 6 percent are upper class. That gives us 12,462,000 potential iPhone users. According to current statistics, iPhones have a global market share of 22 percent, which yields a final guesstimated result of 8,844,000.
8. How many socks would you have to remove from a bag that contains blue and red socks to get a pair of matching socks?
Say you reach in and take out a red sock, then a blue one. Since the third try will give you another red or blue sock, you'll have your pair. So, you need three attempts.
9. How many people live in your apartment building?
Your city has a standardized apartment configuration of 10-floor structures, each floor with 20 apartments. That’s 200 apartments.
The average apartment has four people living in it. So at first glance, your guesstimate would be 800 people. But not so fast! Conventional wisdom says 10 percent of a building’s apartments are unoccupied! So that means you guesstimate that 720 people live in your building.
10. How many cups of coffee do Americans drink in New York City per month?
Let’s begin by establishing that people drink fewer cups of coffee on the weekends because they don’t need the caffeine boost for their jobs. Next, we need to look up the population of NYC, which is about 9 million people.
Let’s now say that 20 percent of these people are children, and you don’t want to caffeinate children! So of the remaining number, 30 percent drink coffee every day, 20 percent drink coffee occasionally, and 10 percent drink tea instead.
Now, we assume that daily coffee drinkers could be drinking three cups of coffee a day, and the occasional coffee drinkers are happy with just two cups a week. Here’s the formula breakdown:
Daily coffee drinkers: 3 x 0.2 x 7 = 4.2
Occasional coffee drinkers: 1 x 0.2 x 1 = 0.2
Tea drinkers: 0
Total: Daily + Occasionally + Tea drinkers = 4.4 cups in a day
Per month = 4 x 4.4 x 7.2 million = 126,720,000 cups of coffee per month. That’s a lot of java!
Looking forward to becoming a Data Scientist? Check out the Data Science Bootcamp Program and get certified today.
Do You Want to Become a Data Scientist?
Our world runs on data, and the need for data scientists is greater than ever. If you would like a career as a data scientist, Simplilearn can help you get your start. The Data Scientist Master’s program, held in collaboration with IBM, boosts your Data Science career aspirations and provides you with the needed world-class training and skills to become successful in the field. The course offers extensive training on the most in-demand Data Science and Machine Learning skills, including hands-on exposure to critical tools and technologies like Python, R, Tableau, and the concepts of Machine Learning. You can become an expert in Data Science by diving deep into the nuances of data interpretation, mastering technologies like Machine Learning, and mastering powerful programming skills designed to take your career in Data Science to the next level.
According to Glassdoor, Data Scientists in India can earn an annual median salary of ₹1,000,000. Glassdoor also reports that Data Scientists in the United States can make a yearly average salary of $110,000.
So don’t delay! Visit Simplilearn today and get started on a rewarding, exciting career in Data Science!