Ask Away - How To Create A Portfolio, A/B Testing, and SQL Interviews
Reposting and answering some of your common questions.
Question:
I am having a struggle with SQL technical interviews. I know SQL and was using it a lot in my previous job. Yet in an interview, I’m feeling lost and can’t get the right answer. Whatever approach I try, it’s either doesn’t work for that question or an interviewer expected something else. Anything you can advise to help me improve SQL?
Oh, I feel you, technical coding interviews can be so challenging. It comes with practice.
Try to get used to a framework of tackling the question:
break it down on subparts
solve each part separately
connect all together
If your code won’t work, at least you could walk your interviewer through your logic.
For example, if the question is to return a list MAU for a specific time period, you can start from (1) defining “active” users, then (2) getting monthly active users, and (3) work on formatting the given date-time range.
Also, try to understand and practice functions. Most likely, you will be asked to apply RANK or SUM or AVR.
Here is a good post on what to focus on in your practice: https://towardsdatascience.com/the-practical-sql-guide-i-wish-i-had-when-i-started-data-analysis
And here is a guide on what type of questions expect on your SQL interviews and how to practice SQL: https://dataanalysis.substack.com/p/sql-hack-sql-interview-questions.
Question:
I was wondering, do you have more practice interview questions for a senior product analyst position? My interview is supposed to check my “statistics competence and how to approach analytical problems in a product context”. If you were to interview someone for this position, what questions would you ask? Right now I’m preparing for questions about A/B testing, but not sure if that’s all there is to prepare.
A product analyst is a data analyst “adapted” for a product lifecycle environment. First of all, they will definitely ask you about A/B and another product testing and will dig down into your understanding of statistics. Be prepared to answer the questions about how long you have to run the test, how to know how big a sample you need, different caveats. Make sure you know how to explain statistical power, significance. They might ask you to define and calculate what statistical test you would use for A/B testing: a t-test or z-test or chi-square. Once I received this question during my interview.
They most likely will ask your understanding of product metrics. How would you define active users? Which top-level metrics you would pick for success measurement? How would you calculate them? Which conversions or rates to focus on?
Be prepared to tackle some specific case scenarios. It can be about how to improve user churn, increase user engagement, or fix user retention. If you are interviewing into revenue team or SaaS, you must know what LTV is, how it is calculated, and how to do revenue forecasting.
But it also can be a position in the Marketing team. Then they will focus on SEO tactics, user engagement metrics, open rates, click-through rates, maybe some CRM applications, and metrics.
Question:
How do I build a compelling data analytics/data science portfolio so that my application to Data Analyst/Data Science roles is a compelling one?
Having a portfolio that illustrates the work you did and your skills is important. Here are my thoughts and recommendations:
I noticed some data analysts add publications or articles as their portfolio on Medium or other blog platforms. Here are some good analysis examples:
The most common tool for a DS portfolio is Github. Here are some examples:
You also can set up GitHub pages. Learn here how to do it.
I am a fan of Kaggle notebooks - easy to start with, easy to share, and add people to collaborate. Here are some Kaggle projects which can work as your portfolio:
https://www.kaggle.com/olgaberezovsky/predicting-titanic-survival-using-most-common-ml
https://www.kaggle.com/jainmegha835/is-handwash-necessary
If you have a question, you can comment here or email me at olga@berezovsky.me.