The editor believes that "realizing data-driven business and becoming a great data scientist" is the dream of many data analysts when they start. However, after learning a lot of statistics and data analysis software, many people have become unfeeling counting machines. Many students are puzzled. What is their current level and how can they be considered as having data analyst skills?
Image source: Design team of Haidata Lab
Saying nothing, the editor came up with a way: as we all know, what can fully reflect the qualities required by the labor market for a certain occupation should be the recruitment of JDs. After all, everyone has worked so hard to improve their abilities for satisfaction. Requirements on JD.
Therefore, Xiaobian drew the following word cloud map by crawling the job requirements data of 500+ data analyst positions on a recruitment website.
Image source: Design team of Haidata Lab
Observing this picture, we can see that, in addition to some of the characteristics required by any profession, such as responsibility, team communication, and fast learning, the job requirements of a data analyst can be simply divided into four categories: understanding business, understanding analysis, understanding tools, Understand the platform.
1 Understand business
Understanding business is the biggest prerequisite for engaging in data analysis.
That is, data analysts should be familiar with industry knowledge, company business and processes, and preferably have their own unique insights. Data is the product of business, and business, in turn, is the real service object of data analysis. If data analysis is divorced from industry knowledge and company business background, the result of the analysis will be like an off-line kite, meaningless.
The business itself has a very broad meaning, divided into five parts: business model, organizational structure, business process, business strategy, and implementation. But to put it simply, an introductory data analyst wants to understand a company's business, just to understand the following questions.
What are the departments in the company? Which department do I connect with?
What are they most concerned about?
What are they doing recently?
Where does the data come from and where is it used?
Image source: Design team of Haidata Lab
These problems are very simple, as simple as holding the docking person for two minutes to solve, but it is also very important. To make data meaningful and affect decision-making, an understanding of business and indicators is essential. Of course, this is also the most lacking ability of many analysts, because this ability not only requires a certain accumulation of knowledge, but also a lot of work experience.
2 Understand analysis
Data analyst positions in the market are usually mixed. Although some positions are called data analysts, they either only deal with Excel every day or deal with code every day. Business analysts are not as good as operations, and tool analysts are not as good as technology. Therefore, a true data analyst should be based on the word "analysis".
Understanding analysis requires data analysts to master the basic principles of statistics and methods to interpret data. Just like living in this world we need to understand the laws of the world, the world of data also has specific laws. If you do not understand statistical knowledge and analytical methods, then the data are just meaningless numbers. Among them, the ability of data analysis thinking and quantitative interpretation are the skills that data analysts rely on for survival.
Taking the decline of active indicators as an example, analysts should address the following issues:
How much did the active indicator fall? When did it start?
Is the data fluctuating within a reasonable range or is it sudden?
Is the overall number of active users falling, or are some of them?
Why did it fall? Is it a product version issue, operational error, or other issues?
How to solve the problem of falling?
If a data analyst finds that user activity in a certain area has dropped after multi-dimensional analysis, he hurriedly submits this as the analysis conclusion report. Then, this analyst might be scolded. Because this is just a phenomenon, it is the problem, not the cause.
The final step of the analysis should be to find out why the activity in this area has declined. Whether it is the local channel, competitors, or market environment, these all need further in-depth quantitative analysis.
Image source: Design team of Haidata Lab
At the same time, finding standards is also crucial for data analysts! Because the data itself does not reflect the problem, data + standard can reflect the quality of the data.
However, newcomers often only judge single indicators with clear meaning standards, such as sales, profit margins, and other standards, and they often judge too superficially and superficially, believing that "a decline in an indicator is not good, but a rise is good."
In fact, a decline is not necessarily bad, and a rise is not necessarily good. It must be judged based on other criteria. For example, due to environmental reasons this year, the sales volume of a certain product dropped a lot compared to the previous few days. This seems to be very bad, but if compared with competing products, the sales volume is much higher than that of products in the same industry. The fall is not entirely bad.
Entry-level data analysts should master the following four single-dimensional standard methods:
The above methods are simple single-dimensional standards, while standards involving two or even three are more advanced requirements for data analysts, requiring more complex dimensionality reduction methods or comprehensive evaluation methods.
3 Understand the tools
Understanding tools refers to mastering common tools related to data analysis. Generally speaking, understanding basic statistics, mastering Excel, SQL and basic visualization is enough to complete most of the work of data analysts, but more advanced data analysts also need to master Python, R and machine learning capabilities.
If understanding the business helps data analysts find problems, and understanding analysis helps them to disassemble the problems, then understanding the tools plays a role in realizing the theory. They are not the most important, but they are indispensable. Especially in today's Internet era where the amount of data is increasing, data analysts cannot rely on human brains and simple calculators to process data, but need to find more advanced and smarter data analysis tools to assist their analysis process.
Image source: Design team of Haidata Lab
However, remember that tools are only an aid, not a fundamental. Cultivating analytical thinking and business understanding is more important for data analysts.
4 Understand the platform
Finally, let's talk about the understanding platform of data analysis. What is called understanding the platform is simply to have platform awareness. When we understand the business, understand the analysis, and understand the tools, we will become a good analyst. We can analyze the project by ourselves, refine the data, and output the business strategy, but if we When we want to save more repetitive manual labor and control more projects, we need to rely on our data platform technology.
Mass access of various types of data, and business also requires fast data feedback, data collection-data cleaning-data modeling-data visualization, this series of processes requires data analysis to make data productized and platformized.
For example, major companies now build their own BI platforms, data science platforms, and data analysis will deposit daily business problems on the data platform, so that both the business and themselves can quickly obtain the required data, perform data indicator modeling, and regularly output Lead cockpit data, or visual data reports, etc.; or directly access various third-party data platforms.
Understanding the platform is a kind of consciousness. It is also one of the very important methods for data analysis to improve the efficiency and value of the routine and tedious work.
5 Conclusion
Image source: Design team of Haidata Lab
In general, if you can understand business, analysis, tools, and platforms, then congratulations, you are already a data analyst in the industry.
In the next step, you have to learn how to independently initiate, be responsible, promote projects, see the results of data analysis, and realize data-driven business. This is also the core symbol of becoming a higher-level data analyst.
Cover·Picture/Feng Wei
Character / 覃 欣 懿
