Shulou(Shulou.com)06/01 Report--

Chatting with a friend A the other day, a data analyst who seems to me to be a company in most industries is still modestly calling himself a "data ape". Engaged in data analysis, he became a monk halfway. Although he knew some knowledge of database at first, he still struggled to get on the road like many people. So we simply discussed the scriptures together and talked about our work "seriously", leaving some insights.

How does data analysis start?

It must be a lie to say that the introduction is very simple. A said that he started by memorizing the data, on the one hand, to cope with the leaders' questions, and on the other hand, to cultivate the sensitivity of the data. Indeed, the source of analysis usually comes from significant changes in some indicators, and familiarity with daily trading data or user data allows you to see at a glance what the problem is and which data are relevant, and then analyze them. I started with Excel and spent most of my time taking numbers, being a "cousin" and resisting the demand from the business. Later, he went to do data mining, and I switched from BI to data platform.

Everything is difficult at the beginning, but once the data analysis has the motivation, it is necessary to improve their own knowledge system, which is also the beginning of the real introduction. So how to improve the knowledge system of data analysis?

1. Basic computer knowledge and statistical knowledge

Database + SQL language

Some commonly used databases such as Oracle, SQL Sever, DB2, MySQL, these databases or daily contact databases should have some understanding, understand the most commonly used, the most important thing is to be able to write SQL.

Mathematics / statistics knowledge

The importance of some basic mathematical statistical methods, such as descriptive statistics, multivariate statistical analysis, regression analysis and so on, is self-evident.

Data mining knowledge: analysis of variance, regression analysis, factor analysis, cluster analysis and so on. These things are more or less used as an introduction, although they may not be used completely, but they are hated when they are used.

Visual tools for data analysis

Data analysis visualization tools are broad. The first is Excel, small and medium-sized companies are very dependent, proficient in the use of PivotTable, which is a necessary skill. Large and medium-sized companies use report tools or BI to make reports, but with the foundation of SQL+Excel, these tools are very quick to get started.

2. Business knowledge

Data analysts have to deal with the business of the company, so they should have an in-depth understanding of the business knowledge of each department. A business leader needs to know a certain indicator, and you need to know what data this indicator consists of? What is the caliber of data statistics? How to get the data out? What is the significance of this indicator to the industry, in what scope corresponds to what kind of situation, whether it is good or bad. Then slowly explore the multi-dimensional law of this index level, how to set the most reasonable.

Know your position and grow up quickly.

Attach a data analyst competency system diagram on the Internet for reference.

Data analysis has always been a relatively professional job, you should always be vigilant about whether your ability has been improved, what your current level is, and get used to reflecting on yourself:

Here is a quote from Zhihu @ Ren Mingyuan's answer

1. Do you know the source of the data you collated? Is it your own company's business data, or is it data exchanged with partners? Is it collected by the relevant departments of your own company, or from a third party? What are the specific indicators and logic in the acquisition process?

2. Are these data true? Will there be any problems in the process of collection and sorting? Technical logic and business logic are different concepts. Are there any data processes that are technically flawless but do not conform to business logic?

3. What has been done with the data in your hands? What did you do with it? Why are they doing this with you?

4. Who needs your data? Where does your processed data flow to? What do they do with the data? What did the data eventually do with it? For example, what services have been done for customers, what content has been released by the company, what KPI has been proved to management, or which department's evaluation has been supported?

5. What is your finishing cycle? Why such a cycle?

6. Is there any other department in the company dealing with other data? What kind of data is it? What's that got to do with you? Why should these data be processed separately?

7. In the past year, you should have accumulated a lot of data on your own computer. Try to do an analysis. What has happened to this piece of data you are responsible for over a long period of time? Why did this change happen? Is it related to the company's products, operations, business, or industry? How exactly is it related?

How to further improve?

In business

1. Business as the core and data as the king

Understand the structure of the whole industry chain

Make a good business development plan

Understand the core indicators of measurement

With the data, it must be combined with the business to be effective.

Need to understand the overall situation of the business, find out the whole structure of the industrial chain, and have a general understanding of the upstream and downstream operation of the industry. Then, according to the current needs of the business, specify the development plan, thus classifying the data that needs to be sorted out. The last step is to list the data core indicators (KPI) in detail, and disassemble several core indicators in more detail, of course, combined with your business attributes to deal with, to find out those factors that have a greater impact on the indicators. The collection of preliminary information and a comprehensive grasp of the current business situation are critical.

two。 Consider the present situation of the index and find the multi-dimensional law.

Be familiar with the product framework and fully define the operational status of each indicator.

Compare with the indicators of the same industry, excavate the hidden room for improvement.

Disassemble the key indicators and set up the operation method reasonably to observe the effect.

Compete for core users, conduct product research and demand mining separately

Most of the business analysis is qualitative, and it is necessary to cultivate an objective sense of consciousness. Qualitative analysis requires the help of technology, tools and machines. As for the cultivation of feeling, because everyone's thinking and perception are different, we can only control the general direction, and the relationship between many data elements still needs to be realized through data visualization technology.

3. Verification of laws and summary of experience

After finding the rule, it can not be online immediately, and the model needs to be verified on the testing machine.

Skillfully

Is 1.Excel a fine diamond?

In addition to the commonly used Excel functions (sum, average, if, countifs, sumifs, offset, match, index, etc.), Excel charts (pie chart, chart, bar chart, radar chart, etc.) and simple analysis skills are also often used to help you quickly analyze business trends and anomalies; in addition, the function inside Excel combined with PivotTable and VBA function is a sharp weapon to improve report development, allowing you to easily complete the report with one click.

two。 You need to know more about databases.

Commonly used databases such as MySQL,Sql Server, Oracle, DB2, MongoDB, etc.; except for the skillful use of SQL statements, we should also be proficient in the storage and reading process of the database. When dealing with a large amount of data, it is very necessary to find ways to speed up the running speed of the program, reduce network traffic and improve the security of the database.

3. Master data collation, visualization and report making

Data collation is to convert the original data into a convenient and practical format. Excel is not a good tool for collaborative work. Report FineReport is more recommended. Project deployment of Tableau, FineBI, Qlikview and other BI tools, there is no good training and learning, these convenient tools can play down some repetitive operations in data analysis, and focus more on analysis.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.