Regardless of being interviewee or interviewer

Being for 7 years on the job market, I’ve been interviewed a couple of times and in the past year, I conducted a few of them myself. I am a data analyst, which meant something different at every company I’ve worked at. Technical interviews for this position can cover a wide range of topics, which might look scary at first. And even if a data analyst can’t be a swiss army knife, we do have experience with many disciplines. In this article you’ll read about my experience and how I would prepare for an interview now.

For this article I asked my colleague, Dorka Benko () to provide her insights as she conducted many interviews within the company. Thank you Dorka for taking the time and effort to improve this piece!

What does a data analyst do?

I guess you’ve already met with a couple of requirements in job descriptions and probably you acquired analyst experience from your studies or work. As interviewers, we appreciate, that everyone comes with a different background and during a technical interview we aim to explore both your strengths and current weaknesses. I’d like to emphasize though, if you are not an expert in a field, that’s not a no-go. The job of the interviewer is to understand what your boundaries are and where you could improve to find the best fitting role.

In our company, applicants receive an assignment, where they are free to choose any dataset, perform analyses and present their findings at the interview. Let me share what skills and knowledge I usually assess.

when it come to data analysis i excel meme Via

1. Primary and secondary research

Depending on the phase or business domain, it might happen that you need to perform a research to provide foundation for the project. Do you have experience with primary research (interviews, questionnaires, workshops), with reviewing existing documentations or published papers and with consolidating that information (called secondary research)? It could be a valuable asset later on.

2. Data exploration

Have you ever met with raw data without any instructions what to develop on top of it? When profiling data, you should summarize information of the initial dataset to better understand what type of data you deal with. Techniques could include analyzing frequency, variation, outliers, correlation of variables etc.

3. ETL (extract, transform, load)

If you worked with data preparation, you must have stumbled into the term of ETL. Here I would be interested in your experience with different data sources, harmonizing data from multiple sources, data cleaning activities, data load types and their efficiency.

4. Business Intelligence & Analytics

Extracting knowledge from data. Isn’t it the most important part of our job? BI is the cherry on top, which encourages you to find the relevant questions and tell a story with data. What type of models did you use so far to discover insights and support decision-making? We are interested in data mining, predictive models and all the other cool stuff you did so far. Be aware though, that using these techniques in your application will imply, you are familiar with the statistics behind, so the interviewers will probably ask about that too.

In business intelligence, the number of tools you may choose from is infinite, but in reality, most of us are well-versed in 3-4 tools/technologies probably. According to the current landscape (Gartner 2020 BI Magic Quadrant), Tableau, Power BI and Qlik are the industry leaders. I personally develop in TIBCO Spotfire during the day and have some fun in Tableau when the sun goes down.

And while working in BI tools is enough in most of the cases, there isn’t always an out of the box solution, so you’ll have to implement a custom development to meet with customer needs. I found coding experience in Python, R, JavaScript and SQL to be quite handy as a data analyst, so again, if you mention any of these programming languages, be prepared to get theoretical and practical questions to assess your current level.

me waiting for my algorithms to run baby yoda meme Image via

5. Business understanding

Once you have done the initial analysis of your dataset, you’ll know which business questions might be relevant, and how those could be answered. Keep your focus on providing business value. As I emphasized above, it is even better, if you tell a story (using examples from your data), which has an analytical flow, starting from high level understanding, guiding the end user through questions, new aspects and drilling down to examine key datapoints in a greater detail. Creating a story from data could be the first step to sell your ideas, so your dashboard should demonstrate your understanding of business needs and serve these needs with every interaction at every level of information.

6. Data visualization

Ok, that might be as important as BI, since it’s not enough to gain insight, but you must share it with your audience. How do you choose visual types? How do you organize the layout of a single page dashboard and how do you handle multiple pages? Is there a meaning associated with the colors you chose and are they consistent? Still, visuals are not just about colors and formats. You should remember that your visuals should broadcast a message: what can you conclude from the data? Fancy visuals look nice on infographics, but some of them may not be the best choice for a business presentation. Keep in mind, that managers are busy and don’t have minutes to figure out what they can read from a chart.

Try to visualize data with the end user in mind: what is the most important information they want to spot? I was always a little disappointed when the customer asked for another table chart in the dashboard, but I’ve learned, this is how they validate data. Find the balance between “traditional” and fresh.

But first of all, demonstrate your knowledge of data visualization best practices. Keep your visualization simple and clean: use colors if they have an added value or consider limiting the palette, set a logical interaction between charts and action buttons, remove unnecessary labels and legend, use relevant wording in titles and description.

7. Presentation

Presentations are sent across the company and to the customers every day. They should be clear and concise. In my opinion, aesthetics is an essential skill for a BI developer. A presentation should tell a story and be thoughtful about reflecting its agenda / business questions. Ideally, text and visual stimuli are in balance, so the receiver can understand your message without further explanation.

Keep in mind, your presentation should be a summary of your work, not the verification of all the legwork you’ve done.

8. Customer facing experience

My company is a project-based B2B company, developing solutions for clients. In many cases, the data analyst has to work strongly together with the customer, so it’s crucial to understand during the interview, whether you’ve already worked in such an environment. Elaborating on requirements, written and verbal communication, expectation management, situational questions (what would you do if the customer says…) could be all in focus.

9. Basic understanding of one domain

You may have experience within industries like finance or life sciences. We understand that industries have many segments applying data analysis, so please share your specialty, e.g. a domain expertise like asset/wealth management or credit risk management. Knowing your interest will help in finding your place within the company. Who knows, maybe you’re the missing piece to kick-off a future project!

10. Understanding of business processes and requirements

As I pointed out at customer experience, a data analyst will often interact with the client. That means, you should systematically collect and organize any information you receive. Interviewers may ask about classic business analysis processes you are familiar with, like stakeholder analysis, requirement elicitation, estimation and prioritization techniques or how would you identify and address risks of a project.

11. Development practices

Just like being a data analyst, the way of development differs from company to company. Our projects usually run according to the agile principles. Thus, we are looking for people, who have a similar mindset, but it is even better, if you have hands-on experience working in agile.

12. Attitude, teamwork & management experience

I didn’t talk about soft skills in much detail, as they are hard to assess in an hour and the impression is quite subjective to the interviewer. Actually, I think your overall behavior within that one hour will best demonstrate your attitude. Analytical way of thinking, problem solving attitude, willingness to learn, patience and respectful communication will all come to the surface.

Finally, if you apply to a senior or lead position, that will require you to share your experience about managing a team or a project. How do you delegate tasks within a team? Do you take responsibility for the mistakes your team made? What do you think makes a team successful and what would be your part in it? No one is born to be a manager though; we all learn from the mistakes we made. You can share them too, but as an interviewer, I would carefully watch the way you articulate these issues and how you mitigated them.

How to prepare and what to include in a test assignment

If I were applying for a position now, I’d try to showcase my experience in the above listed areas. How could you do that?

am i ready for this interview meme Via

First of all, you should choose a dataset from a business domain you are familiar with. Let’s say, if you worked with online marketing specialists for years, you could perform e-commerce website analysis. Choosing so, you’ll be able to form relevant business questions and answer ad-hoc questions more confidently.

For such an assignment, 2-4 business questions are enough. However, they should focus to a pain point which could drive decision-making. As a business owner, which question would trigger you: „How many leads did social media platforms generate?” or „Which Facebook ads did generate more than 70% ROI in the last campaign?”

As a second step, you’d probably explore your dataset, clean the data, apply transformations and create derived variables for specific metrics. In a customer presentation that information might be redundant, but it worth a slide for the interview. It implies, that you are conscious about data and try to bring out the best of it. Which reminds me of a classic: „Garbage in, garbage out.”

In many assignments I’ve seen applicants adding too many charts, just to show us, they could run complex analyses however they forgot to explain why and how the specific visual tackles the overall problem. In my opinion, in such task less could be more. Don’t include models which you don’t find sufficient to answer the business questions or just mention them at the end briefly. At each slide you should build a story to piece together the puzzle, not to confuse the receiver. Also, find a balance between text and visuals; you don’t always have the chance to present your findings in person, your presentation might be distributed offline.

I always appreciate, when applicants add a slide for summarizing their findings and refer to initial business questions, as this is the take-away of the presentation. Also, I find next steps slide useful, but only if it recommends actionable items.

How to handle questions during the interview

The interviewer must prepare with a wide range of questions. As I mentioned, this is not to frighten candidates, but to assess your level of seniority and determine in what role you fit the best.

Be prepared to receive questions regarding your skills in your resume and the content of your presentation. Even if it’s intriguing to include all the online courses and 1-day workshops you participated at in your CV, I’d advise against that. Rather include skills where you have hands-on experience so you can back it up with examples and won’t find yourself in a weird situation of trying to explain something you don’t really know about. Usual example is ML. NLP (natural language processing) is a trendy area and you’ll find tremendous code snippets how to use it in Python. But can you explain what are the underlying statistical techniques? If the answer is yes, you should definitely include it in your CV, but if your answer is not really, I recommend listing it in Interests section instead. This will save you a lot of trouble during the interview too.

When answering the questions, be aware of time! You might feel that you need to explain a concept in great depth to convince the interviewer you know what you are talking about. Instead, try to summarize your answer in 4-5 sentences, and at a technical interview you can expect others to understand the jargon, so don’t hesitate to use domain specific terms rather than high-level and vague explanations.

When you don’t know the answer for some questions, that’s ok too! If you are not sure, whether you understood the question, just ask the interviewer to rephrase it. If you are not familiar with a concept, it’s better to acknowledge what you don’t know, than wasting everyone’s time with playing a guessing game. Interviewers appreciate when people are aware of their boundaries.

What is the conclusion?

Interviews are never easy. Once done, you’ll come back to questions, and think “why didn’t I say this or that”, like in a movie. But I think, it’s possible to minimize the stress you take, if you keep in mind some advices:

  1. Only put skills in your resume, which you are confident enough to demonstrate your advanced knowledge during the interview
  2. Bring an example where you were responsible for a broad spectrum of tasks in end-to-end data management
  3. If you have an assignment, spark your experience in the disciplines listed above
    1. BUT! Always keep the focus on business understanding
    2. Include advanced analysis methods, where you can explain the underlying statistical model
    3. Present your findings based on an analytical flow or use case
    4. Summarize what is the take-away of your analysis
  4. Collect your list of achievements in advance, which will support your explanation for situational questions
  5. If you don’t understand a question, feel free to ask the interviewer to rephrase it
  6. Keep your answers short and to the point
  7. Avoid repeating the same answer put in other words
  8. When describing a process or method, rather use technical terms than giving a vague explanation

I collected a couple of inspiring articles at the bottom, which are focusing on the importance of data storytelling. Technical skills we can acquire from education, soft skills we can improve with practice, but how you channel your creative thinking could make a real difference, and it should get an equal role in development as the other two.

Well, that’s what I had in mind. I hope, you’ve read some useful tips and comforting words between the lines.

What’s your experience with technical interviews? Share your thoughts on Twitter, I’d love to hear your story and opinion too.


  1. The Art of Storytelling in Analytics and Data Science | How to Create Data Stories?
  2. The next chapter in analytics: data storytelling
  3. Data-Driven Storytelling project by Microsoft
  4. Tell a Meaningful Story With Data
  5. The Ultimate Guide: Data Visualization vs. Storytelling with Data
  6. What Great Data Analysts Do — and Why Every Organization Needs Them