当前位置：网站首页>Stop looking! The most complete data analysis strategy of the whole network is here

Stop looking! The most complete data analysis strategy of the whole network is here

2022-06-24 07:08:00 【Yonghong Data Science Institute】

Imagine a scene like this :

The leader said ：“ You go to the building materials market to buy me some accessories .” You run all over the markets under the scorching sun , But the leader asks you ：“ Why this one ？” But you can't answer .

Didn't you try ？ Hard work . But is it effective ？ At least in the eyes of leaders , There is no point in your sweating .

Actually , The same is true of data analysis . When a leader gives you a task , When you're clueless and just trying to gather data , The final result of the work is equally meaningless .

today , We go from beginning to end , Take a good look at the whole process of data analysis .

First step ： There is clear consciousness in the brain

When a leader gives a task , We should keep asking questions in our minds . Everyone who does news knows 5W1H, This analysis method is also applicable to data analysis .

What： What thing 、Where： Where 、When： When 、Who： Who 、Why： What's the cause of the 、How： How to do .

When we put this 6 A thorough analysis of the problems , Naturally, we found the starting point for data collection , Instead of looking for a needle in a haystack of complex data .

At the end of this step , We can define the data analysis process ： The first step is to get the data , The second step is to analyze the data , The third step is to draw a conclusion .

The second step ： Get the data

Facing massive data , Where do we get it ？ How to take it ？ At this step , We should further subdivide it into building data framework and extracting data .

Building a framework is one of the basic skills of data analysis , Whenever a new business need arises , Should think through the framework , This plays a vital role in solving the problem .

Pass the previous step , We got the data , But can these data be used directly ？ Not at all , We also need to do data preprocessing , Dispose of useless data , Get clean and important data for analysis .

The third step ： Analyze the data

At this point , We also need to master enough analytical methods , Today, let's take a look at the common 6 An analytical method .

Classification analysis ： Classification is a basic way of data analysis , According to its characteristics , Data objects can be divided into different parts and types , Further analysis , Can further explore the essence of things .

matrix analysis ： On the basis of matrix diagram , Put different elements in rows and columns , Then at the intersection of rows and columns , Describe the comparison between these factors in terms of quantity , Then calculate the quantity 、 quantitative analysis , Determine which factors are most important .

Funnel analysis ： The essence of analyzing a problem is to disassemble it , Break down complex problems into small ones . Funnel analysis is a set of process analysis , It is mainly applicable to long processes , There are many links , And the traffic is gradually lost along with the link .

Correlation analysis ： Analyze the signs that do have connections in the whole , Its main body is the analysis of the sign of causality in the whole . It is the process of describing the closeness of the relationship between objective things and expressing it with appropriate statistical indicators .

Logical tree analysis ： Think of a known problem as a tree trunk , Then start to think about which related issues or sub tasks this problem has to do with . Every thought of , Just add one to this question “ The branches of the ”, And mark this “ The branches of the ” On behalf of what problem .

Trend analysis ： Through the analysis of the change trend of each period of relevant indicators to the base period , Find problems in it , An analytical method that provides clues for tracing and checking accounts , The trend analysis method can be used in either relative or absolute numbers .

Step four ： Come to the conclusion

The data analysis has reached this stage , It's the end 、 Reporting stage , That is to conclude . In the past work , We may take Excel Form report work . But there is a very big problem ： Can not find the primary and secondary and pain points .

therefore , Now most data analysts use visual reports to report their work . When making reports , You should use histogram as much as possible 、 The pie chart 、 Line chart, etc , In this way, the results of data analysis can be displayed intuitively , Instead of reporting with your own subjective feelings .

Visual reports can be made by Yonghong Desktop, This product is not only very friendly to novices , The full function is provided for free forever , No matter who is just beginning to learn data analysis , Or a professional data analyst , It can be used to make visual reports to report work .

Yonghong Desktop Built in multiple theme styles

Since then , Through the highest bar in the histogram 、 The section of the graph with the greatest fluctuation , We can see at a glance where the problem is .

Yes, of course , It is not enough to have handy tools , Making visual reports is an art , It doesn't just require you to have strong logic , Know how to position charts , We should also have good aesthetics , From color matching 、 Font, etc , Highly ornamental 、 Visual report with strong practicability .

Step five ： Learning in the second round

Data analysis requires long-term and in-depth learning , With the right tools , We should also master enough 、 A methodology that is effective enough . Next , Recommend some good things to everyone .

Books ：

Introduction to data analysis ：

《 Data analysis 》

《 Who says rookies can't analyze data 》

《 Naked Statistics 》

Advanced stage of data analysis ：

《 Master web analytics 2.0》

《 Website analysis practice 》

《 On statistics 》

《 Data management 》

《SQL From entry to mastery 》

《 Data operation manual ： Method 、 Tools 、 Case study 》