7 Data Analytics Interview Questions & Answers - KDnuggets

• Read original article here

The data analytics interviews are divided into multiple parts, such as non-technical, technical, and SQL. The hiring manager will assess your knowledge of the statistical tools and concepts. Furthermore, you will be asked situational questions where you have to explain how you prepared an analytical report, cleaned the data, or came up with graph interpretation. 

In this blog, we will go through 7 questions that are changeling and frequently asked during the data analytics interview. 

In this question, the interviewer is judging your communication, presentation, and people skills. Being able to explain technical concepts to managers or clients is a skill. 

Apart from technical terms such as mean, correlation, or data distribution, you also need to learn more about data and its features. Try to connect dots that make sense for a business. You need to make sure you understand the business and audience to explain concepts in layman's terms. 

To answer this question, you need domain knowledge of industry, business, and the product. You can ask an interviewer to tell you about the company strategy and vision, which help you formulate the answer. 

For the social media product, the 3 metrics can be daily active users, number of users adding friends in the first 2 weeks, and the number of posts in a week. It is based on the company's vision and product strategy. So, it is always better to research the company before sitting for an interview. 

Descriptive analytics provides insights into the past to answer the question such as “how did the marketing camping perform compared to last year”

Predictive analytics is about using insight to predict future events or forecast growth. 

Prescriptive analytics is used to suggest various courses of action to prevent disaster or to improve the product. 

This question is totally up to you. In general, data analytical projects consist of understanding the problem statement, gathering the data, cleaning the data, exploring, analyzing, and visualizing the data, and finally interpreting the results for the non-technical audience. You can also mention tools, techniques, and additional steps based on specific problems. 

There are various ways to handle missing data. The most used method is dropping the missing values rows if the dataset is large and balanced. 

Apart from that, you can:

The solution is simple. You will select the required columns and Count(*). After that, group it by unique identification, such as employee name, manager id, joining date, and city. We will then use HAVING to filter duplicates. If the Count(*) value is greater than one, then it is a duplicate. 

You can apply the same strategy to any table. Make sure you are grouping tables by multiple unique id columns such as name and address. 

The solution is simple but tricky. First, you have to count the number of images per user and then count the number of users with more than 1000 images but less than 2000. 

The inner query will count event_date_time and group it by user_id to find a unique user id with a number of images per user. After that, create an outer query to filter out users with more than 1000 but less than 2000 images and count them.