What is business intelligence (BI), just look at this article is enough

01、 What is business intelligence BI？

business intelligence BI - Parker data business intelligence BI Visual analysis platform

First, understand what business intelligence is BI（ Business Intelligence ）？ Baidu business intelligence BI There are many, many official definitions , Various interpretations , In fact, from so many years of experience , For business intelligence BI I don't think it needs to be complicated , Just three ：

First of all , business intelligence BI Is a complete set of Data warehouse 、 Query report 、 Data analysis And so on Data technology solutions .

second , business intelligence BI You can integrate different business systems of an enterprise （ ERP、OA、CRM etc. ） The data in Get through And carry out effective Integrate . This connection and integration includes ：ETL The process 、 Count 、 Business logic rules are like the transformation of data rules 、 Data warehouse modeling, etc .

Third , business intelligence BI Finally, we can use appropriate query and analysis tools to provide visual analysis or reports quickly and accurately , Provide... For the enterprise Decision support .

  business intelligence BI - Parker data business intelligence BI Visual analysis platform

business intelligence BI Do you need to explain it so tall ？ Unwanted , In fact, it is simple to see the third floor ：

first floor , Visual analysis presentation layer  - Business intelligence BI The requirements layer of , Represent the needs of users , What users want to see , What to analyze is displayed on this layer .

The second floor , Data model layer  - Business intelligence BI Data warehouse layer , An analytical model that represents data , Complete the transformation from business calculation rules to data calculation rules .

The third level , The data source layer  - Business intelligence BI The data layer of , The data of the underlying database of each business system passes through ETL To extract to business intelligence BI In the data warehouse ETL The process , Modeling analysis, etc , Finally, the visual analysis display supported to the front end .

02、 business intelligence BI In Enterprises IT Position in informatization

This point is that if all enterprises plan to use business intelligence BI The project must be understood ： business intelligence BI stay IT What is the position of informatization ？ Clear positioning is an important prerequisite for information planning and construction .

Usually , We will be planning business intelligence BI Before the project , Put the enterprise's IT Informatization is divided into two stages ： One is business informatization , One is data informatization . In this way , It is easier for the average user to understand .

Enterprise informatization - Parker data business intelligence BI Visual analysis platform

Business informatization  - Enterprises use ERP、CRM、OA、MES、 Self built business system, etc , The construction of business systems is collectively referred to as business informatization . The main function of business informatization is to manage the business process of an enterprise , Standardization 、 Online , To improve production and operation efficiency 、 Reduce enterprise cost 、 For business intelligence BI Lay a data foundation for the construction of 、 It is the embodiment of business management ideas , It is also a modern way of enterprise management .

Data informatization  - As we often hear about big data 、 business intelligence BI 、 Data analysis 、 Data mining and so on are collectively referred to as data informatization . Data informatization can help enterprises comprehensively understand the operation and management of enterprises , From experience driven to data driven , Form business decision support , To improve the accuracy of decision-making , This is a higher level of enterprise management .

There is no business system construction , There will be no precipitation of data , There is no business intelligence BI The basis of . meanwhile , business intelligence BI The construction of can reverse the construction of business informatization .

Enterprise informatization - Parker data business intelligence BI Visual analysis platform

The main use forms of business informatization  - Tabular 、 Mainly business user input 、 Most of the data are added, deleted and modified , Business process data 、 A software system that manages business processes .

The main use forms of data informatization  - For example, business intelligence BI It mainly presents the overall analysis and local insight of the business result data , It aims to get through the data of various business systems , Cross business 、 Integrating data across systems . Mainly focus on data query and analysis , Through linkage 、 Drilling 、 Data indicators can be viewed in the way of graph visualization such as correlation .

03、 Who is business intelligence BI Primary users of ？

  business intelligence BI - Parker data business intelligence BI Visual analysis platform

The main users of business informatization  - Front line business executive layer , More from a business perspective , Input data 、 Record the process 、 View business information .

The main users of data informatization  - Management and decision making , It is more about business intelligence from the perspective of management BI Visual analysis to locate the problem 、 To analyze problems , Finally, business decisions are formed .

Two details ：

First of all , There is no management decision-making level 、 The leaders will open the financial system to see the financial data , open OA The system looks at the contract information , Senior leaders won't look at these details , I won't go into various systems to see . in other words , Business informatization is not for leaders at this level .

second , Does the management decision-making level necessarily refer to the top leaders of the enterprise , Not necessarily , It can be the personnel with management decision attributes in all organizational levels of the enterprise , These management decision makers can use business intelligence BI Provide decision support .

04、 What does data island mean ？

Why do we have to use business intelligence to eliminate data islands BI, Who wants to eliminate the data island ？ Does the business department feel that there is a problem of data island ？ I use my financial system to keep accounts , Data island is just an island , I like isolated islands , My financial data is for us to look at , Leaders' View , I am not alone at all . I'll keep an inventory , Data island is just an island , I don't care about anything else , My report is good enough to read , What does the island have to do with me ？

therefore , We are talking about business intelligence BI, When talking about data islands, it is not for front-line business departments , It should be for cross business 、 Cross department 、 These management levels across organizations say , Only from their perspective , These business systems and data are the real islands .

business intelligence BI - Parker data business intelligence BI Visual analysis platform

What is the underlying reason ？ The underlying reason is ： These business information systems do not serve the management , It serves the front-line business department . Management is not the user of these business systems , They will never log in system by system to see the data , They don't have this habit of using , They don't pay attention to the micro level of each business system . therefore , In most cases, only these Cross organizations 、 Cross business management will consider the existence of data islands , So they asked to solve the data island .

Due to business intelligence BI It is a natural solution to the problem of data islands , So business intelligence BI For whom , It is a data information system serving the management . business intelligence BI To break the data island , Look at the data comprehensively , Comprehensive management of business , business intelligence BI It is a natural extension of the business management perspective , Want breadth 、 Want depth .

  Manage cockpit - Parker data business intelligence BI Visual analysis platform

therefore , From different angles , Some people think that there is a data island , It must be solved . Some people do not think that there are data islands , Even their existence has no effect on them , So there's no need to solve , The fundamental reason is that they do not grasp business intelligence BI Real service objects .

Through data islands , We can see some problems more clearly .

05、 business intelligence BI The method of fetching data from the business system

business intelligence BI It is not like the interface between the business system and the business system to develop data retrieval methods , It accesses data by accessing and connecting the data source database of the business system , No matter what type of database it is , business intelligence BI adopt ETL Connection mode connect the database to extract the original table data of the business system and process it in the data warehouse , Visual analysis report presentation supported to the front end .

business intelligence BI - Parker data business intelligence BI Visual analysis platform

A friend asked this question before ： Does the data source layer need to develop interfaces ？

This is the answer ：

In general, you don't need , Basically, those who ask such questions have experienced interface docking of software systems , The interface connection of software system is because some business software is JAVA Developed , There are plenty of them .NET Developed , There are plenty of them B/S framework , There are plenty of them C/S framework . The interfaces between software systems need to be developed , It mainly connects the business processes of different software , This interface requires dynamic code . but business intelligence BI The interface for obtaining data is different , It has nothing to do with the business system software itself , You only need to access and connect the database behind the business system , Get data directly from the database , Therefore, no software interface is required , Or there is no software interface to access this concept .

Except in one case , This business system is a public cloud , pure SAAS Pattern , In this case, it can only be opened through software API The interface fetches data .

Kanban of a construction industry group  - Parker data business intelligence BI Visual analysis platform

  After sales business analysis of an automobile industry - Parker data business intelligence BI Visual analysis platform

06、 Data center 、 business intelligence BI、 How to understand the relationship between big data ？

We are learning about business intelligence BI When , It is inevitable to learn about big data , Some concepts such as data center , How to correctly understand the differences between them ？ big data 、 Both data centers are business intelligence BI A product developed to a certain stage , The core is all around data , Data collection 、 Data processing power 、 The improvement of computing power gave birth to big data , Data assets and data services give birth to the data middle office , The data warehouse modeling of the core is from business intelligence BI It comes down in one continuous line and has not changed , The final export is business intelligence BI visualization , So business intelligence BI Is at the top of the whole information construction .

07、 About business intelligence BI Several cognitive mistakes

About business intelligence BI Introduction to , There is too much noise on the network , In a word, business intelligence BI The explanation seems very simple , I feel that buying a tool can solve all the problems , This is actually a very big misunderstanding .

  business intelligence BI - Parker data business intelligence BI Visual analysis platform

I've summed it up here , Everyone is interested in business intelligence BI Some misunderstandings often encountered in the understanding of ：

1. business intelligence BI Report visualization , It's just a bunch of visual charts , business intelligence BI Front end visualization .

2. business intelligence BI It is a drag and drop analysis tool product .

3. business intelligence BI It's business intelligence BI, It has nothing to do with the data warehouse .

4. With business intelligence BI There is no need for data warehouse modeling , Business people can do business intelligence by themselves BI analysis , You can drag and drop to do business intelligence BI analysis .

5. business intelligence BI It is business driven , Unwanted IT Personnel support , Agile Business Intelligence BI Unwanted IT intervention .

6. business intelligence BI Isn't the direct connection fragrant ？ Directly connecting to the data source is not enough for analysis , No need for data warehouse .

First, briefly correct the understanding of these problems .

1. business intelligence BI Report visualization , It's just a bunch of visual charts ,BI Front end visualization .

business intelligence BI Is a complete set of data warehouse 、 Data analysis 、 Data report and other data technology solutions , In a BI In the project ,20% Time for front-end Analysis Report ,80% The time is spent in the design of the underlying data warehouse 、ETL Development of 、 Access development, etc .

So visual reports are just business intelligence BI The final presentation of , But it's not business intelligence BI All .

2. business intelligence BI It is a drag and drop analysis tool product .

The drag and drop visual analysis tool can only solve business intelligence BI Part of , Visual analysis . But in fact business intelligence BI The scope of technology covered is quite wide , It involves all aspects from low-level data retrieval to front-end presentation analysis . Take Microsoft for example , As early as SQL Server 2005 You can see the complete Integration Service（ SSIS ）、Reporting Service （ SSRS ）、Analysis Service （ SSAS ）, These three services plus SQL Server It constitutes Microsoft's business intelligence BI Solution .（ SQL Server 2000 Of DTS I don't know how many people remember ）Reporting - Visual presentation is just business intelligence BI Part of the solution .

SSIS To solve something  - ETL Tools ,Extraction extract - Transformation transformation - Loading load , Do the whole ETL Visual design of 、 Package management 、ETL Package scheduling management , Contains Package - Control Flow - Data Flow Manage the whole data extraction . Data warehouse DW The layered design of the system , for example ODS / Staging layer 、Dimension layer 、Fact layer （ from DW To DM ） The logical table data extraction of layer can also be put into SSIS Done in .

I was Microsoft business intelligence BI Technology line , I have accumulated a lot in the early years business intelligence BI Technology blogs - BIWORK Technology blog

SSRS To solve something  - Reporting Report presentation , At the beginning, the presentation of the statements was relatively weak . stay 2012 Windows 8 Metro UI When the design came out , We are SSRS We learn from it Metro UI The style of , It's more avantgarde and amazing .

To be fair , Many report tools at home and abroad have been used for reference SSRS Data set schema for （ Write a SQL A query or stored procedure returns the result of a query bound to a data set Dataset in , Charts are bound to datasets , The fields of the chart are referenced from the dataset ）, However, this method also has its restricted use scenarios or use prerequisites , This question will be discussed later .

SSAS To solve something  - The multidimensional analysis of space for time ,OLAP、CUBE Cube . For example, multiple dimensions in an analysis report ( Dimension ) Can be combined with multiple metrics （ Measure ) Combine , In time 、 Area 、 Take the three dimensions of product and the measurement of sales revenue as an example , When the user opens a report , The query that can be combined according to the fields of the report is ：

SELECT Time ,

                     Area ,

                     product ,

                    SUM ( Sales revenue ）AS income

             FROM Fact table JOIN Time dimension table ON XXXXX

                  JOIN Area dimension table ON XXXXX

                  JOIN Product dimension table ON XXXXX

             GROUP BY Time 、 Area 、 product

It may be such a query

SELECT Time ,

                     Area ,

                    SUM ( Sales revenue ）AS income

             FROM Fact table JOIN Time dimension table ON XXXXX

                  JOIN Area dimension table ON XXXXX

             GROUP BY Time 、 Area

When the amount of data in the underlying data table is too large 、 Aggregate queries and complex , A combination of various dimensions and fact measures SQL A large number of queries are sent to the data warehouse , This kind of query efficiency may become very poor , Because data query SQL It may take a long time to execute , This does not include the intermediate data transmission process returning to the front-end report 、 Rendering time of front-end reports, etc , So pass SSAS Achieve one CUBE Cube , The essence is the aggregation query of various dimensions and measures （ Various aggregate functions , You can choose ） SQL It was executed ahead of time , Finally, various dimensions and measures SQL The value of the query is stored in advance . Front end reports are connected to CUBE You can directly use the pre calculated value in , Instead of going through SQL Go to the data warehouse layer to query , This is the principle of space for time .

What problem does that illustrate here , It's a complete set of business intelligence BI It actually includes a lot of things , With underlying data processing ETL The process , There are also front-end visual analysis reports .

stay ETL Tool level ： Microsoft SSIS、Informatica、IBM DataStage、Pentaho、Kettle、DataWatch wait .

stay report form Reporting Tool level ： Early Microsoft SSRS、IBM Cognos、Oracle BIEE、SAP BO wait .

Simple drag and drop BI Strictly speaking, visual analysis tools can only be positioned at the individual and department level business intelligence BI Analysis tools , Because the last one business intelligence BI Analysis tools cannot solve business intelligence BI All , Nor can it replace business intelligence BI All .

3. Some people used to say business intelligence BI Business driven , business intelligence BI Namely BI, It has nothing to do with the data warehouse .

With business intelligence BI There is no need for data warehouse modeling , Business people can do it by themselves business intelligence BI analysis , You can drag and drop business intelligence BI analysis ,, Unwanted IT Personnel support , agile business intelligence BI Unwanted IT intervention , There is no need to build a data warehouse , I used to think so for a while . But after a while , This methodology has been tracked for some time , Finally, it was found that there was a big problem .

  Management cockpit large screen - Parker data business intelligence BI Visual analysis platform

Whenever there is any business intelligence BI Sales or tell users before sales , Of your enterprise business intelligence BI The project does not need to build a data warehouse , Directly through business intelligence BI The analysis tool can be dragged to complete all the analysis in the enterprise , Unwanted IT Personnel support , Business people can do it by themselves ... Similar to those who dare to make such a commitment , Either right or wrong. business intelligence BI Don't understand, , Or it's really fooling .

At the enterprise level business intelligence BI The project is under construction , It can be realized by simply dragging and dropping the data by the business personnel , At least in my personal work experience of more than ten years ,95% None of the above enterprises can . The key enterprises I have served include ：SHP（ Security Health Plan )、 Microsoft （ China ）、 Microsoft （ The United States ）、VWFC（ Popular finance ） etc. .

VWFC It is very good , Few business people make many reports by themselves , There are thousands of reports online . Why? ？ Because the underlying data warehouse has been built for many years , The underlying data architecture is relatively standard .Business Driven Business driven , What is its premise ？

1） The underlying data quality is very standard , The data warehouse architecture is complete , Don't let the business personnel touch the underlying data ,ETL、 Count 、 Index calculation and so on are all IT Department to maintain .

2） Business personnel should master business intelligence through training BI Use of front-end reporting tools , Understand the interface of the released data analysis model .

3） Business personnel should be very familiar with business and data .

The first 2） And the 3） Many enterprises have no problem , The first 1） Get a straight front end business intelligence BI Tools for business people to solve , Can you solve it ？ Obviously, business personnel do not have this ability .

This is the time for training , business intelligence BI The tools are easy to use , But once the actual enterprise business intelligence BI Project development found that it was difficult to move forward . Because when training , The data tables given are all selected , It is always of high quality 、 The standard only needs a simple left table and a right table, such as the sales order table 、 Order details , Naturally, it is easy to implement visual reports .

  Data visualization - Parker data business intelligence BI Visual analysis platform

But in actual enterprises business intelligence BI Project analysis , The calculation rules of analysis indicators can not be solved by simply associating several tables , If you don't believe it, you can challenge a practical indicator calculation logic ： Challenge one ETL A small case of data cleaning There is only one data table in the database , The data is easy to understand , But many business intelligence BI Developers need a lot of energy to do it , Let alone the self-help of business personnel business intelligence BI Analysis of the .

I don't say so much just to deny the self-help business intelligence BI Its function and ability , Self help business intelligence BI There are usage scenarios for it , And it does help us simplify a lot BI Work , But from a professional point of view , A particular antipathy is part of business intelligence BI In an irresponsible way, manufacturers repeatedly reinforce such concepts to the market ： business intelligence BI Visual reports 、 business intelligence BI No data warehouse modeling required 、 Traditional data warehouse modeling is very backward 、 business intelligence BI Is self-help analysis 、 business intelligence BI Self help analysis is simple 、 Business users can learn and analyze as they like after a few days of training ...

From the perspective of marketing and sales , There is no problem in simplifying the publicity of the complexity and difficulty of the product , The problem is with a wrong explanation 、 Unprofessional explanations eventually mislead enterprises into accepting these incorrect concepts , And use these incorrect concepts to evaluate and plan business intelligence BI Construction of the project , Not fully anticipated business intelligence BI Challenges and risks that may be encountered during project construction , Finally, it leads to the failure of the project 、 Repeated construction .

We had a client in Beijing before It cost more than a million On a set of so-called business intelligence BI project , The project has been online for about a year , In the end, it can't be pushed at all , failed . Find Parker data later , We gave them Parker data business intelligence BI Analysis platform , We have been working on this project for several consecutive periods , The customer also wrote a thank you letter . Why couldn't I push it before 、 The project will fail ： Do not pay attention to data warehouse planning . Because their business is continuous 、 Variable , Annual demand needs dynamic adjustment , Data continues to increase , The depth and breadth of analysis are constantly changing , There is no good underlying data architecture to support , The light on SQL Count 、 It is impossible to support the future of an enterprise in the form of data sets and reports 3-5 Business analysis needs change for years or even longer .

In addition to this case , On my mobile phone, there are many people who have done it before business intelligence BI The ultimate failure 、 Didn't do a good job , Find the record of chatting and roast , Is the product really bad ？ I also objectively helped them analyze ： Some of these products are Gartner Magic phenomenon Leader Quadrant products , Do you think the product is OK ？ Some products are domestic business intelligence BI An old brand in the field for many years , Do you think the product is OK ？ To be fair , These products from my personal point of view , These products are actually excellent , The product itself is not a big problem .

The problem lies in , So many from zero to one need to go up business intelligence BI Of enterprises do not know one business intelligence BI There were so many holes in the project , quite a lot business intelligence BI Will the manufacturer explain these points clearly to enterprise customers , One business intelligence BI How to do the project 、 What kind of risks are there 、 What kind of problems will you encounter in the future 、 How to solve these problems 、 What kind of methodology and means ... If only to sell a set business intelligence BI Products or tools , You think these business intelligence BI Will the salesperson tell the customer about these things ？ I won't , At least I won't talk too deeply and completely , Because such a talk put business intelligence BI The difficulty is too complicated , Once you don't speak well , On the contrary, it reduces the trust of customers .

Sometimes I don't say , It's because I'm afraid it's complicated , The decision-making cycle of enterprise customers is too long . Sometimes I don't say , Because I don't understand . You don't say , The customer doesn't know , The customer has no experience , Follow up business intelligence BI There will be problems in project construction .

At a conference , A business intelligence BI A senior pre-sales technical expert of the manufacturer said something when communicating with customers ： business intelligence BI Isn't the direct connection fragrant ？ Directly connecting to the data source is not enough for analysis , No need for data warehouse . Ignorance is bold , I really can't listen , I just interrupted the direct communication . Through communication , It can be judged that this so-called technical expert has not done a complete business intelligence BI Project experience , Build one from zero to one business intelligence BI The capacity of the project is equal to zero . With such a kind of ability to guide a customer business intelligence BI project , such business intelligence BI Can the quality of the project be guaranteed , It's hard .

This is our Parker data 、 I personally make a video number 《 Lvpinliao data 》 Why , Speak objectively business intelligence BI、 Talk about the data objectively , Popularize what we think is right business intelligence BI Knowledge and concepts . Tell us about the vast business intelligence BI user , business intelligence BI How to understand 、 How to recognize , business intelligence BI What kind of pit does our enterprise need to pay attention to .

We can't say that our Parker data is business intelligence BI The knowledge and concepts in the field must be universally applicable , But we welcome any business intelligence BI Manufacturer or any BI Personal enthusiasts are business intelligence BI Some knowledge and concepts to challenge us , Take a look at some of the popular data of Parker business intelligence BI Is the concept of knowledge right . If the popularity is right , It shows that these problems are indeed encountered by everyone , These knowledge and concepts are rare experiences for enterprises . If the popularity is wrong , No, where is it , Point it out and let's have a look , Let's talk about , What else can we do for the enterprise .

08、 How do report tools come from ？

I have been in the field of technology for more than ten years 、 Information technology 、 business intelligence BI industry , I haven't been out of this circle . done JAVA ( AWT、SWING、JSP、Hibernate、Spring、ibatis )、.NET ( ASP、http://ASP.NET、C#.NET )、Object-C 、JS And so on , Business software system platform development .

The early front-end technology was weak ,AJAX The implementation of also needs handwriting , To realize the click editing and modification of data in a form, you need to use it yourself JS DOM operation . Making reports is basically JSP、ASP Scripting languages are nested in the front end HTML Do cyclic output , The report style is very primitive and ugly , A slightly more complex table report style requires JS To adjust .

The reports used at that time were like Crystal Report Crystal Report 、 Moistening and drying reports, etc , There are tags in the front-end scripting language that can directly reference , Report generation replaces a lot of handwritten code . The early front end and back end technologies were not separated ,http://ASP.NET It's a little better , The front end gradually has some integrated controls that can be used directly ,JAVA It's true that there is no . When is the stage mentioned above ,2005 Around the year ,2007 I think it has been widely used in , old CSDN You should also find many original report label posts on the .

Like the old batch of statements and like the Jinfeng statements Jreport、 Star report StyleReport There is also a certain market in China . As early as 2010 Years ago , The revenue scale of some report manufacturers has exceeded 100 million , It shows that the basic report market is still very good .

What was the report positioning at that time , It's pure Report report form , Query the returned data aggregation from the background database through the program List Bind to the front-end script page to generate various reports , In fact, it is the report display used in various business software systems , It's far from business intelligence BI Analyze this level .

And there are a large number of software developers who actually have strong reporting capabilities , However, these reporting capabilities have not been independently used as reporting products to operate in the market .

Gradual , With front-end technology 、 The improvement of the front-end framework , From the traditional table technology to various histogram 、 Bar chart 、 Visual display of pie chart , At this stage , Reporting and business intelligence BI The boundaries of the world are getting blurred . Why? ？ business intelligence BI The report presentation ability of the is roughly the same as that of the traditional report , There is no such thing as self-help analysis 、 Self service drag and drop can realize the ability of rapid multi-dimensional analysis .

What I want to say is a lot of business intelligence we have seen BI Projects are implemented with the report thinking , Namely SQL To data set and front-end presentation . And true business intelligence BI What should thinking be ？ Multidimensional thinking 、 Model thinking , This determines a business intelligence BI The final trend of the project , These points will be discussed in detail later .

09、 business intelligence BI The essence of - The landing of enterprise business management thinking

  business intelligence BI The essence of - Parker data business intelligence BI Visual analysis platform

business intelligence BI What is it ？ technology ？ product ？ Or else? ？ Let's treat BI To raise our understanding of ： business intelligence BI It is the landing of an enterprise's business and management thinking . How to understand this ？ Simply speaking , The content presented on the visual report is the content that an enterprise really cares about , There are business analysis indicators that the top management focuses on , There are also some specific departments .

10、 business intelligence BI And data warehouse Data Warehouse What are the differences and connections ？

People often ask about business intelligence BI And data warehouse , In fact, this question can reflect some friends' understanding of business intelligence BI There are still some inaccuracies and deviations , This problem actually conceptually refers to BI And the data warehouse are artificially separated . This situation is also quite normal , Because everyone is interested in business intelligence BI My first impression is all kinds of cool visual charts 、 report form , In addition, there are many lightweight front-end visual business intelligence in the market BI Analysis tools , This will cause everyone to be right BI The cognition of is just in the part of visualization .

To be exact , business intelligence BI Not just front-end visual analysis 、 The ability of report presentation , It also includes the construction process of the underlying data warehouse .Gartner Business intelligence was mentioned in the 1990s Business Intelligence, It thinks more ：BI It is a technical solution for data class , Many data from different enterprise business systems are extracted and cleaned 、 Transform and load , Is to extract Extraction、 transformation Transformation、 load Loading Of ETL The process , Eventually merged into a data warehouse , According to certain modeling methods, such as Inmon Of 3NF modeling 、Kimball Or a hybrid architecture model of both , Finally, on this basis, appropriate analysis and presentation tools are used to form various visual analysis reports to provide data decision support for the management decision-making layer of the enterprise .

business intelligence BI - Parker data business intelligence BI Visual analysis platform

therefore , You can see the data warehouse from here Data Warehouse Is the layer between the visual report and the data source of the underlying business system , Throughout business intelligence BI The project solution plays a connecting role . If business intelligence BI Compared to a person , The upper body, especially the face, is the appearance , The lower body is down-to-earth and absorbs the essence of the earth , The core of the waist and abdomen in the middle 、 The core strength is data warehouse .

Jordan, the God of basketball, is not only handsome , The ability to stay in the air is a top-notch existence , There will be all kinds of amazing actions in the layup , What actually supports these movements ？ It's Jordan's core strength .

therefore , business intelligence BI At the front-end visual analysis level, we should play all kinds of wonderful actions , Without the support of the core strength of data warehouse, it is very difficult to do .

Then people will also ask , There are not many business intelligence that can be analyzed by dragging and dropping directly linked data sources in the market BI Tool products , No, you can also do business intelligence BI Analysis report ？ This kind of independent 、 Separate front-end oriented business intelligence BI Analysis tools , They are more positioned as department level and individual level business intelligence BI Analysis tools , For deep-seated complex data processing 、 Integrate 、 Modeling and many other scenarios cannot be solved . The best way is to build a complete data warehouse at the bottom , Standardize many analytical models , Reuse these front-end business intelligence BI Combine analysis tools , Only in this way can the front-end business intelligence be truly BI The analytical power is released .

Many enterprises think that they only need to buy a front-end business intelligence BI Analysis tools can solve the problem of enterprise level business intelligence BI All questions , This view is actually not feasible . It may be relatively simple to analyze the scenario at the beginning , When the complexity of docking data is not very high, this kind of business intelligence BI There is no problem with the analysis tool . But in enterprise business intelligence BI The project construction has one characteristic , It is a spiral construction process . Because there may be more and more business systems connected , The depth and breadth of analysis will be more and more , The complexity of data will become more and more challenging , At this time, there is no good data warehouse architecture support , The front end alone BI Analysis tools are basically impossible .

Data warehouse - Parker data business intelligence BI Visual analysis platform

It's like going to a traditional Chinese medicine store to get some medicine , The reason why the medicine is filled quickly , Because before filling the medicine , Others have put all kinds of native Chinese medicinal materials （ Data from the original data source ） Clean it up and put it away , So how to match the herbs （ Visualization of dimension indicator combination ） Soon .

There are many such enterprises in China , It's also about business intelligence BI The lack of depth of understanding leads to a lack of understanding in business intelligence BI Some directional mistakes in project construction , Finally, it leads to business intelligence BI The project is difficult to continue .

So in the enterprise , We need to be clear about our business intelligence BI Is the construction oriented to the enterprise level or the analysis of individuals and departments . If you are a personal data analyst , Use this kind of front-end business intelligence BI Analysis tools are enough . If it is necessary to build an enterprise level business intelligence BI project , You can't just focus on the front-end visual analysis capability , More attention should be paid to the construction of the underlying data architecture , That is, data warehouse .

11、 Modeling methodology of data warehouse Kimball vs Inmon And hybrid architecture

Business intelligence in data warehouse modeling BI The most important thing in the project construction ,Inmon Three paradigms of 3NF Modeling and Kimball The dimensional modeling of is business intelligence BI Methodology of data warehouse modeling , These two kinds of business intelligence BI What are the differences and connections between modeling methods .

12、 Actually carry out a BI Methodology for the implementation of requirements during the project

business intelligence BI Is a completely demand driven , Since it is a demand, we need to do interviews and research .

13、 What kind of enterprises should have business intelligence BI 了 ？

What kind of enterprise is suitable for business intelligence BI？ Look at the level of basic business informatization and the level of detail and granularity of daily business management . The informatization degree of business foundation is the enterprise's own IT Business system infrastructure , Without the support of business system , Do business intelligence BI There is no data base ; The second is the granularity of business management , Is the business management level of the enterprise more meticulous , Business intelligence is urgently needed BI To improve business management 、 Efficiency of decision support .

14、 How to effectively do for senior leaders BI Summary of data analysis report

Finish business intelligence BI project , Also consider how to report to the boss in the end , Master business intelligence BI Data analysis thinking framework and five key points of report ： User business level and scope 、 Results of work 、 Plan to execute a double disk 、 Problem feedback 、 Vision planning and vision .

15、 business intelligence BI Integration with enterprise management

business intelligence BI The analysis is highly combined with the business management analysis of the enterprise ,ROE High profit enterprises may be as high as Maotai 、 Jewelry industry , It may be that the turnover is fast, for example, in the retail industry , It is also possible that the financing ability is relatively strong and leverage will be used , from ROE Attribution analysis looks at the characteristics of the industry .

16、 business intelligence BI Accumulation of project industry and business knowledge

Do business intelligence BI You must also be familiar with industry and business knowledge , Not combined with industry business knowledge , business intelligence BI It is difficult to implement the project .

17、 About business intelligence BI The topic of real-time processing

business intelligence BI There is a certain lag in data processing , Usually used T+1 Pattern , The main reason is ETL The data processing process needs a lot of time consumption , It is usually a way of exchanging space for time .

According to business intelligence BI Data warehouse is hierarchical ETL Scheduling is designed to automatically find dependent scheduling according to individual indicators, which greatly increases the flexibility of individual indicator scheduling and quasi real-time processing .

Offline data processing and real-time processing are different in business scenarios , The technology behind the implementation is different , Resource input is also different , Understanding the positioning differences between them is helpful to select the appropriate solution to achieve the business intelligence of the enterprise with the minimum resource investment BI The goal of the project is .