A fan shared his interview experience:
1. What are the common interview questions?
Mainly prepared from the following knowledge points: common analysis methods, Excel, SQL, A/B testing, product analysis.
Then each interview is based on the job requirements, and after chatting with HR in the early stage to understand the position a little bit, I will prepare for orientation.
The preparation of Excel and SQL is actually learned from the monkey teacher. The course is quite good, suitable for those who have no foundation or weak foundation to check for omissions and make up for vacancies.
Before the interview, it is recommended to review the "Illustrated SQL Interview Questions" again, and the problem will not be too big. For example, in the interview, I encountered several different solutions to the classic "select the top three in each group" (this question is too classic, and it can help you familiarize yourself with the various uses of join, window function, and where)
The A/B test interview will test some basic statistical concepts, such as how to determine P value, type 1 error, normal distribution, and sample size.
Generally, you may ask how long does the test take to run? Why use a t-test instead of a z-test, have you considered the data distribution? Should the experiment be stopped once the results are significant? Do you need to observe the experiment all the time?
How to deal with social effects (A and B's circle of friends are social, A's use of new features will affect B), and novelty effects? The result is one high and one low, how to make a decision?
A/B test results are not significant, what decision do you make? How would you do an experiment if you had three or four instead of two groups? What problem should we pay attention to?
The product analysis is mainly to read 40 product questions in a DS data challenge.
In fact, AB testing is also a part of product analysis. Generally, you will be asked, what software do you usually like to use, talk about what you like and what you don’t like, and how to improve it? Then it will lead to an AB test question, asking about indicator selection, sample data, target users and so on.
The following is the mind map I summarized, which is not comprehensive and can be used as a reference for answering questions
The classic decline problem: What indicators have declined recently, and can you find out the reason?
I have temporarily learned that there are three analysis frameworks: no matter what kind, I personally like the first simple and clear
internal reasons (product function changes, product activities, technical problems, etc.) vs external reasons (time, place, competing products, advertising platforms) etc.), this kind of interview questions have been mentioned in the 10 common analysis methods course of Mr. Monkey, so I won’t repeat them.
2. Talk about the interview experience
The interview process I went through was as follows:
1) HR chat ----> a homework (SQL+ statistical questions or analysis report) (may have this, most of them don't)
2) The manager chats about the projects on the resume, as well as basic analysis or statistical knowledge (see what department the manager is in, and decide whether the direction of the chat is product or statistical knowledge)
3) SQL test, on-site interview ranging from two to three and a half hours (with the product manager, with the team members, and finally chat with the manager and HR)
Generally, HR will chat first, and will ask what they have done before, what they are doing now, do you know our company, do you know this position, why the company, and what position. Salary requirements.
For technical and analysis questions, managers generally ask questions about the projects on their resumes. I have been asked: commonly used business indicators such as ARPPU and other indicators.
There is also the reason for the decline in analysis indicators. If you want to increase your income, what indicators should you look at? If you want to improve the revenue mechanism of the game, what indicators need to be paid attention to.
What should I do if the two indicators of the AB test result are high and the other low? AB test results are not significant, what is your decision?
After the experiment, the idea of a certain advertisement is better than others, what should I do at this time? How to advertise online?
How to do user classification? What strategies do different paying users adopt?
The above is my job search experience sharing, I hope it can help you.
⬇️Click "Read the original text"
Sign up for free Data analysis training camp