What to prepare for a data analyst interview

1.Curriculum vitae

Everyone knows that you must bring a resume to an interview, so how can you create a resume that satisfies the interviewer? Here we suggest that you try the star rule, which can highlight your achievements in data analysis projects.

In addition, the resume must be made in combination with the recruitment requirements, the higher the matching degree with the recruitment requirements, the easier it is to be found by HR, don't be lazy, use a resume to win the world.

2.Delivery

It is best not to submit your resume by sea. If you like one company, you can choose multi-platform delivery.

3.Interviews

Finally, it's time for the most critical part. The advantage of introducing the projects you have been exposed to in general is that there is room for the interviewer to ask questions based on your introduction, and if you say it in too much detail, the interviewer may ask some in-depth questions, and it will be embarrassing if you can't answer them.

There will be technical questions in the data analysis interview, Excel + SQL + Python R These tools are required, and we must master the theory and practice of these tools in detail. The purpose of data analysis is to promote the business growth of the enterprise, and everyone should also know more about the business aspects of the company.

Regarding what to prepare for the data analyst interview, Qingteng will share it with you here. If you have a strong interest in big data engineering, I hope this article can be helpful to you. If you want to learn more about the skills and materials of data analysts and big data engineers, you can click on other articles on this site to learn.

What is a Data Analyst Certificate?

Hello is mainly different depending on the business you are designing. Here's a share of someone else's work experience.

I graduated in July '18 with a bachelor's degree in applied statistics. My first job was as a data analyst in a central enterprise in Nanjing, and my main business was to do some customized data analysis for the public security. When I was looking for this job, my situation was probably like this, I failed the graduate school entrance examination, I didn't have a decent internship experience, I studied a little machine learning course on my own, and my city of work was Nanjing.

At that time, I had a solid foundation in mathematics and probability theory, but I lacked practical work experience, so I didn't have many opportunities to choose. In addition, I have no concept of career planning, and I didn't think too much about finding a job as a data analyst that seemed to be the right one. Before I officially joined the company, I started to read some courses on machine learning and data analysis, including watermelon books and statistical learning methods.

In my first job, I traveled a lot, and I was sent to go to Party A's site to complete the project alone as soon as I joined the company, which was really stressful. Then I started to study hard outside of work, and my career goal at that time was algorithm engineer, because I felt that I was good at math and made a lot of money. So I began to learn frantically according to the skill needs of algorithm engineers.

The courses I have read are probably Python, computer networking, machine learning, data structures, big data, natural language processing, of course, not every one of them is very in-depth. During this period, I also participated in some data science competitions in Tianchi and Kaggle, mainly to learn how to model. However, due to the limited amount that can be used in the work, and more importantly, there is no professional team to lead it.

Because this job is basically staying at the customer's site, just one or two colleagues, all the problems and growth encountered in the work have to rely on their own continuous learning, and overall there is still no growth. Considering the barriers of this industry and my future development, I decided to enter the Internet finance industry, so I resigned at the end of September 2019 and went to Shanghai, the magic capital.

What is data analytics for?

Collecting, calculating, and making data available to other departments in the enterprise.

What is data analysis used for?

From a workflow perspective, there are at least 5 types of analysis that are often done:

Planning analysis before the start of work: to analyze what is worth doing before the start of the work**type analysis: ** the current trend, the expected effect of the monitoring analysis in the work:

Monitor the trend of indicators and find the cause of the problem: analyze the cause of the problem and find the countermeasure after the review analysis: accumulate experience and summarize lessons.

Please click Enter a description.

So what is data analysis?

Data analysis is broadly divided into 3 steps:

1: Get data. Obtain user behavior data through burying points, and open up internal system data through data synchronization. and the construction of data warehouses to store data.

2: Calculate data. According to the analysis requirements, extract the required data, calculate the data, and make tables.

3: Interpret the data. Interpret the meaning of the data and derive some useful conclusions for the business.

So do data analysts mainly do the above three things?

It's not all, this is different in different companies. If the company is large, the acquisition of data is often done by the data development team, and their position is usually "data development engineer" or "big data engineer". Interpreting data is to write your own PPT for interpretation, leaving it to the "data analyst", which is actually a step in the middle of calculating data.

Some companies (generally doing e-commerce), the data is directly exported from platforms such as **, Tmall, Amazon, etc., and then analyzed based on these data. In some companies (generally traditional enterprises), the data is directly used in large-scale BI products, and then everyone exports data analysis based on BI products, and some companies are very small, so they directly do everything from data burying to data warehouse to data withdrawal.

Please click Enter a description.

1. How to understand overfitting?

Overfitting, like underfitting, is a fundamental concept of data mining. Overfitting is when the data is trained too well and errors can occur in a real-world test environment, so proper pruning is also important for data mining algorithms.

Underfitting means that the machine learning is not sufficient, and the data sample is too small for the machine to form a self-awareness.

2. Why is Naive Bayes "naïve"?

Naive Bayes is a simple but extremely powerful modeling algorithm. It is called Naive Bayes because it assumes that each input variable is independent. It's a tough assumption, and it's not necessarily true, but the technology is still very effective for the vast majority of complex problems.

3. What is the most important idea of SVM?

The process of SVM computation is the process of helping us find the hyperplane, and it has a core concept called classification intervals. The goal of the SVM is to find the hyperplane of the value that is the largest of all the classification intervals.

Mathematically, this is a convex optimization problem. Similarly, we divide SVMs into hard-spaced SVMs, soft-spaced SVMs, and nonlinear SVMs based on whether the data is linear or not.

4. What is the difference between k-means and knn algorithms?

First of all, these two algorithms solve two types of problems in data mining. K-means is the clustering algorithm and KN is the classification algorithm. Secondly, these two algorithms are two different ways of learning.

k-means is unsupervised learning, i.e., we don't need to give a classification label beforehand, while knn is supervised learning, which requires us to give a classification label for the training data. Finally, the meaning of the k-value is different. The k value in k-means represents the k class.

The k value in the knn represents k nearest neighbors.

1. Company welfare.

A lot of the company's sincerity is not only reflected in wages, but also in real economic benefits. Including commercial insurance, meal allowance, provident fund ratio, these are real money.

2. Clear overtime situation.

Don't ask directly about this, but be sure to ask. My general practice is to ask HR to increase the overtime as appropriate under the current expectation after presenting the salary expectation.

In this way, he can basically set up a more realistic overtime time.

3. Performance evaluation criteria.

This is also something that is really about self-interest. Whether the enterprise has KPIs, whether there are OKRs, and how to manage them are all realized in individual performance evaluation.

Many part-time workers don't know the way of goal management, but let me tell you, the difference in goal management plans directly affects the comfort of your work. For example, if a data analyst evaluates performance from the perspective of how much is needed, then he is a position that takes numbers and does not assume the function of analysis.

The reporting relationship will affect your position in the company, do you have a mentor, do you report directly to the leader, and who gives you performance? Who is the leader of your leader, and how is it in the overall organizational structure?

The organization in a large company is very complex, and if you don't ask about this relationship, you may end up in a small department to do chores.

4. The values of the direct leader.

I asked him about his understanding of data analytics and his plans for the future of the organization. And the purpose of recruiting me, as well as the understanding of the functions of the number, to have an in-depth exchange.

This way I can confirm that this person is available to work with, and not like other companies interviewing with.