Data analysis children's shoes can't be missed | operational risk control business analysis report

For small partners engaged in credit data analysis , Sorting out and producing business-related analysis reports is a very familiar content in daily work , For example, strategy analysis 、 Model monitoring 、 Customer portrait 、 Post loan analysis and other topics . however , A good credit data analysis report , Not only for sample data and business needs , Use effective analysis ideas to summarize the core content points , And according to different analysis results , Display through appropriate visual charts , This can make the final analysis report highly readable .
Chart generation for data analysis report , We often use excel Tool implementation , Although in many cases, you can easily get the desired results , But through python After analyzing the sample data, it is also necessary to transfer the relevant data into excel in , This sometimes seems superfluous , And if python It is more convenient to implement the language environment directly , among matplotlib The relevant functions of the library provide a convenient way for the generation of data analysis charts .
In order to illustrate the expression effect of diversified charts in the data analysis report , Through a sample data of actual credit products , Common chart types are used to analyze and describe the characteristic distribution of the customer group applying for loan , So as to build a customer portrait data analysis report .
Let's first understand the example sample data , It comes from the basic information of the pre loan customer application of a credit consumer product , Include 6000 Samples and 7 Features , front 10 The observed data of samples are shown in the figure 1 Shown , The corresponding feature dictionary is shown in the figure 2 Shown
​
chart 1 Sample data

​

chart 2 Feature dictionary

Next, we pass the above sample 6 Characteristic dimensions , To describe the distribution information of customer groups . Of course , In the actual scene, feature cross analysis can be used , The specific needs depend on the business needs . This article describes the basic information of customers for convenience , And the display of relevant visual charts , Only the single feature dimension is analyzed . For the characteristic variables analyzed , Due to fields age The value of is relatively scattered , And in the actual scene, the age analysis is often described by age group is more appropriate , Therefore, it is transformed in the form of discrete interval , The implementation process and value results are shown in the figure 3 Shown .
​
chart 3 Feature conversion

Analysis dimension for single feature , We go through python Language to analyze the distribution information of each characteristic variable in turn , The specific chart forms include histogram 、 Broken line diagram 、 Bar chart 、 Scatter plot 、 The pie chart 、 Area map , These visual graphic methods are also often used in our data analysis and report sorting tasks .
1、 Age range （ Bar charts ）
Histogram is used to analyze and describe the characteristic dimension of customer group “ Age range ” The distribution of information , The specific implementation process is shown in the figure 4 Shown , The visualization results are shown in the figure 5 Shown .
​
chart 4 Age range analysis

​

chart 5 Age range display

From the above results , The age distribution of the customer group is mainly 25_{40 Range , In especial 25}35 The interval is relatively concentrated and the proportion is obviously high , And age 45+ The above groups account for less .

2、 Housing type （ Area map ）
The area map is used to analyze and describe the characteristic dimensions of the customer group “ Housing type ” The distribution of information , The specific implementation process is shown in the figure 6 Shown , The visualization results are shown in the figure 7 Shown .
​
chart 6 Housing type analysis

​

chart 7 Housing type display

From the above results , The housing type of the customer group is “ Self purchase without mortgage ” The largest number , This shows that the asset capacity of the customer group is good , and “ Self owned mortgage ” And “ Rent a house ” The situation is relatively more , This is also in line with the actual scene .

3、 Education level （ Broken line diagram ）
Use line chart to analyze and describe the characteristic dimension of customer group “ Education level ” The distribution of information , The specific implementation process is shown in the figure 8 Shown , The visualization results are shown in the figure 9 Shown .
​
chart 8 Education level analysis

​

chart 9 Education level display

From the above results , The education level of the customer group is mainly “ Specialty ” And “ Undergraduate ” Two types of , And “ Specialty ” The number of clients with academic qualifications is the largest , On the whole, it also reflects that the education level of the customer group performs well , And for “ master ”、“ Doctor and above ” Higher education , Although the number is relatively low , But it is also completely in line with the actual situation .

4、 Applied quota （ Scatter plot ）
Scatter diagram is used to analyze and describe the characteristic dimension of customer group “ Applied quota ” The distribution of information , The specific implementation process is shown in the figure 10 Shown , The visualization results are shown in the figure 11 Shown .
​

chart 10 Apply for quota analysis

​

chart 11 Application quota display

From the above results , The application limit of the customer group is in 6000 Mostly , And the number of customers varies greatly compared with other quotas , Then the limit value 8000、3000、4000 More , The proportion of other customer groups with quota value is relatively close , From here, we can also see the general scope of applying for customers' capital needs .

5、 Channel type （ The pie chart ）
Use pie chart to analyze and describe the characteristic dimension of customer group “ Channel type ” The distribution of information , The specific implementation process is shown in the figure 12 Shown , The visualization results are shown in the figure 13 Shown .
​
chart 12 Channel type analysis

​

chart 13 Channel type display

From the above results , The main incoming channels of customer groups are CH01、Android（ Natural flow ）, The number of these two types of customers accounts for about 65% The proportion of , And for channels ios（ Natural flow ）、CH03、CH04 The number of customers in other forms is relatively low , This distribution facilitates the overall analysis of flow sources .

6、 Registered address （ Bar chart ）
Use bar graph to analyze and describe the characteristic dimension of customer group “ Registered address ” The distribution of information , The specific implementation process is shown in the figure 14 Shown , The visualization results are shown in the figure 15 Shown .
​
chart 14 Registered residence address analysis

​

chart 15 Registered residence address display

From the above results , The application geographical distribution of customer groups , Top of the list in quantity 3 The provinces and cities are Guangdong 、 jiangsu 、 Zhejiang , And the least number 3 The provinces and cities are Tibet 、 xinjiang 、 ningxia , From here, it can also be reflected that the difference of regional economic strength has a great relationship with the capital demand of the customer group . meanwhile , According to the regional distribution data of customer groups , It can effectively take certain risk control Approval Measures for the application scope .
The above content is based on an actual sample data of credit , Use a variety of chart forms （ Histogram 、 Broken line diagram 、 Bar chart 、 Scatter plot 、 The pie chart 、 Area map ）, The age of the sample customer group in turn 、 Housing type 、 Education level 、 Applied quota 、 Channel type 、 The registered residence address is 6 Four information dimensions are analyzed and described . From it, we can not only know the distribution information of the customer portrait of the sample group , Moreover, the readability of data analysis report is improved through the application of multi class charts . In the real world , It can integrate business needs and sample conditions , Take more characteristic dimensions to analyze the portrait information and marketing value of customer groups , At the same time, more diversified charts are used to show the content of data analysis .
In order to make you more familiar with the above customer portrait , Sample data and python Code , For your reference , For details, please move to the knowledge planet to view the relevant content .
​

…

~ Original article