This is a big proposition, and it is difficult to make it clear in one or two sentences, so it is difficult to get a satisfactory answer to this question in QQ group.
Here, I will talk about my learning methods as a data practitioner. Of course, one thing to say is that everyone's ideas, methods, work experience and knowledge focus are different, so how to learn this problem can be said to be different. Let me just talk about my personal methods here, which may not be correct or applicable to everyone.
The positions of data analysis can be said to be very diverse, ranging from data entry clerks to industry analysts and experts, and even some people engaged in data mining and artificial intelligence can be included in the scope of data analysis, but what they do is far from the same, and of course the treatment is also very different. Therefore, when applying for a job, don't just look at the title. It is important to look at the responsibilities and requirements of the post. Anyway, let's talk about how to learn data analysis.
I. Knowledge and skills 1. Subject knowledge: According to the professional knowledge points involved in data analysis, it includes many subjects, including but not limited to the following:
(1) Statistics: parametric test, nonparametric test, regression analysis, etc.
(2) Mathematics: linear algebra, calculus, etc.
(3) Sociology: mainly some quantitative statistical knowledge of sociology, such as questionnaire survey and statistical analysis; There is also some sociological knowledge, which is helpful to data analysts engaged in marketing.
(4) Economic Finance: If you are a data analyst in this industry, economic and financial knowledge is necessary, so I won't say much here.
(5) Computer: People engaged in data analysis must know how the data you use is processed and understand the structure and basic principles of the database. At the same time, if the conditions are enough, you can still have enough ability to extract the data you need from the database (such as using SQL to query). This ability to extract data and analyze raw materials is necessary for every data practitioner. In addition, if you want to go further, you must master some programming skills, so that you can borrow some professional data analysis tools to help you finish your work.
These professional knowledge can't be fully mastered in a short time. The only shortcut to learning is reading books, watching videos, reading authoritative books and reading comprehensive knowledge. There is no quick way to learn basic knowledge, because it will be boring and long to learn for basic reasons. How do you hope to have a long-term development in data analysis? I hope you can persist in learning basic knowledge for a long time.
2. Software operation: What are the necessary tools for data analysis? I roughly list the following categories:
(1) Analysis report category: Microsoft office software (excel, word, powerpoint, visio…… ...), xcelsius, etc. If you can't even handle the basic operations of excel forms and even PPT reports, then I have to say that you are still far from the post of data analysis.
(2) Professional data analysis software: OFFICE is not all. To do a good job in data analysis, you must be able to use (at least understand) some commonly used professional data analysis software tools, such as SPSS, SAS, R, Matlab and so on. These softwares can help us to complete professional algorithm or model analysis.
(3) Auxiliary tools: such as MindManager (such as MindManager and MindMapper). ) can also help us to clarify our analytical thinking.
What needs to be explained here is that software is just a tool to help us complete the task. It doesn't mean that you can complete the task well as long as you learn the software operation well, because how to explain the final result is much more important than the operation. Even if the software operation is familiar, if you don't know the result, it's no different from doing it without knowing it. And seeing the results requires solid professional knowledge.
For the above two points, what route to learn and how to arrange the order, I have seen a picture on the Internet, which I personally think is very good:
3, industry knowledge and work experience: how to say this part of knowledge, if you can't learn it in books, it is also deceptive, but what you can really use for yourself is mostly learned in the actual work process. Doing data analysis must be closely related to the industry you are engaged in, and data analysis without combining business is tantamount to an armchair strategist. There are countless industries that need data analysis. In short, as long as there is data, data analysis is needed, such as the Internet, e-commerce, finance, telecommunications, manufacturing, retail and so on. These are all great demands for data analysis. You can't understand every industry, but you can understand it in one industry, and this understanding needs to be accumulated slowly in the course of work.
Second, talk about the relationship between the three. Make an image metaphor. Being an elite data analyst is like being a martial arts expert (many friends should have seen martial arts movies). Wulin experts usually have three elements: vigorous internal skills, deadly moves/rare weapons, and Jianghu experience.
Basic knowledge and experience in the industry, just like this profound internal strength. Even if you do nothing, you can guarantee that others will not cheat you, because you are already in the industry;
Various software operations are like deadly moves and rare weapons. Once shot, it can be fatal and get twice the result with half the effort.
Walking in the Jianghu, I am most afraid of the lack of Jianghu experience. Sometimes I am killed and I don't know who killed me, so my work experience is just like this Jianghu experience. When I have rich experience, it is much easier to deal with problems.
Therefore, the three complement each other, and any short board will affect the overall play and personal data analysis ability.
Third, talk about how to study 1 and read books.
In my opinion, the best way to master knowledge comprehensively and systematically is reading. When reading, you can only read the right books, not the wrong ones. Choosing a book that can greatly improve your ability and thinking is to read the right book. Again, I won't recommend books. There are many classic books in each article, but I can tell you a good way to find books, which is to search the corresponding keywords in the online bookstore. For example, if you want to find books on statistics, then if you want to read books on EXCEL, you can search "Statistics" and "EXCEL". You will find many related books. You can check the catalogue introduction and related evaluation of books to see if they are suitable for you.
2. Visit professional websites
The other is to visit some data analysis forums and blogs frequently. The so-called shopping, just like shopping, I can go shopping without anything. So even if you don't want to find a solution to the problem, you may want to go shopping, because there are a lot of knowledge and opinions about data analysis, and many contents may benefit you a lot. You can also pay attention to Master Daniel and some trends in the industry.
3. Learn to seek answers from search engines.
A person who knows how to learn must be a person who knows how to ask questions, so where is the person who answers your questions, not in reality, but on the internet. When you encounter a difficult problem, I suggest you find a book at hand to help you solve it first. If not, please search on Google and Baidu. Nine times out of ten, the answers to many questions can be found online (of course, those answers are not necessarily the best). If you can't find the answer, well, I admit that your question is a bit biased, so go to the relevant QQ group or ask your colleagues and friends around you.
In addition, in software operation, learn to think about the answers in the operation manual.
Many books about software tools only write the main operation methods. For individuals, the use of a software is only a small part of the function, but the software operation manual is different. It is the instruction manual of the software, and every detailed function point will be written in it. It can be said to be the most complete software dictionary, and almost all the operation methods can be found in the operation manual.
Why do you arrange the order like this? In my opinion, the answers in books are more reliable than those on the Internet. This kind of reliability does not mean that there are no good answers on the Internet, but that you have no ability to distinguish which answer is the best. On the other hand, books are different. The author's knowledge is usually higher than the book he has written. Although the answers given in the book are not necessarily the best, they are not too bad.
Why put the search engine in the second place?
Because search engines can find almost all the contents of the whole network, in a word, everything they find. Learning to use search to find answers to questions is an ability and a method.
If you can't find the above methods, you can only ask friends and netizens for help.
Why QQ group is not a good way to solve problems (except for some flexible problems)?
First, there are experts in the group, but experts are usually busy. They will be happy to help you if they can answer you in a word or two. If not, they usually keep silent. Second, although there are experts in the group, there are also many novices. It is better not to ask than to get a wrong result.
You may have to ask what QQ group is for, and my answer is: solve the problem of flexibility, exchange learning experience and understand the dynamics of others.
Ask friends and colleagues around you for advice, in line with the principle that asking for help is not as good as asking for help. If a friend is enthusiastic and knows the answer himself, he will tell you, and even if he doesn't know, he will sometimes help you find a way. One more thing, asking friends for advice can often play a role in communicating feelings. But one thing, everyone is very busy at work, so it's best not to disturb others.
In short, learning is a step-by-step process, and it is important to persist and not to rush for success; Because data analysis involves a wide range of contents, the principle of learning should be set in a wide direction, and then the knowledge should be continuously expanded and deepened. "Where can't it be supplemented?"
I have written so much without giving you any knowledge, which is my personal opinion and experience. Please correct me if there is anything wrong.
This article is reproduced from data control, link: /p=27.