First of all, business intelligence and report automation .

business intelligence , also called BI, Using modern data warehouse technology 、 Online analytical processing technology 、 Data mining and data presentation technology for data analysis to achieve business value .

Report Automation It refers to the traditional manual collation of reports （excel） To achieve automation of the process , For example, calculate the total sales of the month , The automatic report can automatically calculate the total sales information of the current month according to all the sales order records of the month , It doesn't have to be done manually like the traditional way - statistics .

business intelligence ≠ Report Automation , It can be said that report automation is a foundation of business intelligence , Only by automating the processing of a large amount of data first 、 Summary 、 Statistics can further practice business intelligence .

As mentioned above, report automation is the foundation of business intelligence , So what else is needed ？ We can definitely know that there must be data mining , But this step is relatively complicated , In the process of gradually realizing business intelligence , What we are looking forward to is to generate value continuously in the process of realization , I think it can be divided into the following steps ：

Report Automation

Since reporting automation is the foundation of business intelligence , So implement it first , What is the value of this step ？

• Release the labor force ： When the data is so huge 、 Miscellaneous times , Any enterprise will have all kinds of data , Relatively simple statistics may only need to analyze one type of data , But a little bit more complicated will result in the fusion of multiple types of data or complex analysis on the time line , At this point, a lot of manpower is needed , The realization of automation can directly reduce the repeated development of this part , Note that “ repeat ”.
• Reduce the error rate ： Automation means persisting processes through coding 、 Logic , When we debug it right once , So we can trust this tool , I believe that the stable output rate of machines is more reliable than that of human beings in the repeated work scenario .
• High timeliness ： Except for the monthly reports , We can also consider the implementation of statistics on each day of the month , Collect daily information in real time to dynamically control the monthly plan .

Data charting

The pie chart 、 Broken line diagram 、 Histogram …… I'm talking about charting here , It's not a one-step Visualization , Even visualizing the level of the big screen .

Why do you have a chart first ？ It's just a small step , But it can definitely improve the utilization of data to a great extent , The graph can help us visualize initially .

Preliminary data understanding method ： Why curing ？

A number is placed on the screen independently , What is the connection between them ？ Who and who needs aggregate analysis ？ Charts can be used to solidify business needs or data experts' sensitivity to data , The relationship between the data they found is persisted into code and displayed in a fixed chart , So that any user can see the data and the relationships between them .

Data visualization

What's the relationship between charting and Visualization ？ In fact, the representation of the above diagram is preliminary visualization , But one or two charts can be quickly understood , Dozens of them ？ Too much information requires classification , We can do that .

1、 Graph aggregation

We need to deal with the relationship between charts reasonably , Although a chart can show multiple data or even multiple dimensions of information , But there is a limit to what a chart can represent , We may need to combine multiple charts to show , The relationship between this combination may be ：

• According to the business mix ： Go on to the supermarket , We can according to the sales situation 、 Upstream suppliers 、 Employee management and other different businesses are classified 、 grouping , Combine reports , Display this type of report in a visual area for comprehensive analysis
• According to the value mix ： There is a need to break down departmental barriers in management , It's the same with data visualization , Supermarket purchase and sales will have a certain relationship , The speed with which some items are sold may be affected by the number of salesmen on the day 、 The impact of the number of customers entering the store on the same day , According to a certain relationship, the data that may affect each other / Calculate the formula for statistics , Get valuable information , For example, performance indicators such as conversion rate
• According to the information combination ： Some information may need to be brought together for display , For example, supermarket items that are about to expire

2、 Display optimization

When we put all kinds of information together and show it , We also need to beautify the data

• When it comes to multi-dimensional, we can directly see the evaluation score of each dimension in the same stage in the form of polygon ;
• Information such as ratios can be displayed in colored bars or circles , In addition to being able to see the ratio, you can also directly alert people by color

In the end, we get “ Visualization screen ”, For example, the effect in the figure below , It's all by FineReport It's done .

Through the report we can see some trends , For human visual observation , We may be able to see more clearly the trend shown by the line chart , But how to quantify trends ？

Other forms of reporting , Even the rules among the data that are not organized into reports ？

Data mining can be done at this time , Find potential connections between data in a variety of analytical ways , Making data more valuable .

Last

There must be an end

This article mainly talks about how to gradually achieve business intelligence BI, Through reporting Automation 、 Data charting 、 Data visualization 、 Four steps of data mining , Step by step, let the data generate value .

Of course, this is just the beginning of this series of articles , I'd rather say the first step , Report automation how to do , The following articles will explain my understanding of report automation step by step 、 A design method of report system 、 Data warehouse dimensional modeling process and my practical experience .