Two Bi development, more than 3000 reports? How to do it?

This business intelligence BI The project is probably from  2013 Started in , To 2015 In, it has been supported by more than a dozen business departments ,200 Multiple users , The amount of data is about 370GB about , launched 2000 Multiple business intelligence BI report form . To about 2017、18 In the year , The amount of data will probably grow to TB about , The report grows to 3000 Multiple sheets .

A business intelligence like this BI project , How much business intelligence is needed BI Developer input ？ Some people say , Business intelligence of at least ten or dozens of people every year BI The development team , After all, there are so many departments , So many business people , So many reports .

business intelligence BI Project history

Let me say something about , I was in 2014 I entered this business intelligence around the year BI Project , The whole business intelligence BI The development team is actually four people , Two place  BI Development 、 a BA、 A project manager , So there are only two real developers , I happen to be one of them . Then you must have questions , How can these thousands of statements depend on so many people . Like our own business intelligence BI project , Each business person raises a demand every month , It's dozens of hundreds of statements , It's impossible to do it .

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

Such is the case , Most business intelligence BI The project is difficult to achieve this level . Let me say something about , How did this project achieve . This business intelligence BI The project development is roughly divided into several stages ： First stage , Implementation of data warehouse . Almost from 2013 In, it took more than a year to build a complete data warehouse platform , Of course, the follow-up has also been continuous iterative development . Second stage , User self-service reports . Through business intelligence BI Tool training , Let business users develop reports by themselves .

business intelligence BI Project resumption

This business intelligence BI Now let's see what level the project has reached , At least I didn't feel much when I was in the project , Because there is no comparison , I think the project itself should be done like this . But later I saw a lot of business intelligence BI project , Looking back, I think this project was very successful at that time . There are mainly the following factors ：

1、 business intelligence BI Project index management

First of all , This business intelligence BI In the project ,IT And business boundaries are clearly divided .IT Be responsible for the design of underlying data architecture , The business department makes its own statements , And it is not allowed to touch the underlying data . Any business intelligence required by the business BI The analysis indicators must be unified as a whole , Business personnel are not allowed to create conceptual indicators by themselves . Maintenance of all indicators 、 Development 、 Logical updates must be made by  IT Department to complete .

  Standardized management of indicators  - Parker data business intelligence BI Visual analysis platform

Do business intelligence BI Although the project sacrifices a certain flexibility of front-end business , But it ensures that no matter how many business users , See a certain indicator , Their concepts are exactly the same . How does the underlying data come from , How to design and develop indicators , This is IT Department business . therefore , Even business intelligence BI The system maintains hundreds of indicators , It can also ensure that all indicators do not conflict , Definitions and concepts are highly consistent .

We also have a lot of business intelligence now BI The way to do this is to let the business staff create their own indicators , This method does have some flexibility . But once business intelligence BI The amount of indicators in the project has come up , There will be indicators you build and indicators I build, although the names are the same , But there are completely two different meanings . For enterprise level business intelligence BI Development , Especially later maintenance , Basically disastrous .

2、  business intelligence BI Project data warehouse

second , For business intelligence BI Attention to the construction of the underlying data warehouse of the project . Make great efforts to sort out the underlying data architecture , Demure 、 Be patient to get through all business systems , Sort out business relations . according to  Kimball The way of dimensional modeling builds a complete star model 、 Snowflake model architecture , Layering is also reasonable , from ODS To TRANS layer , Until then DIMENSION dimension 、FACT Fact layer, etc .

  Data warehouse  - Parker data business intelligence BI Visual analysis platform

When this model is polished relatively universally , When scalability is enough , What we open to business department personnel is business intelligence BI The analysis model . Business personnel can only see the analysis models one by one . For example, the required dimensions or indicators cannot be found in this model , It doesn't matter. , We will maintain it after fully communicating and understanding the requirements . This ensures business intelligence BI Dimensions in the project 、 indicators 、 Consistency of the model 、 A high degree of unity .

therefore , What is business intelligence BI At the heart of , The model is business intelligence BI At the heart of . Do we care what the final report looks like ？  did not . We don't even know what kind of reports users will make , Because the following thousands of reports are made by users . Each user has dozens less , There are hundreds of reports , They are all designed and developed by themselves .

3、 business intelligence BI Project business talents

Third , This business intelligence BI The ability and quality of the business personnel involved in the project are very strong .  I remember very clearly , Once when communicating a business logic with a beautiful woman , She wrote a data formula on the paper . I was surprised after reading it , Rarely do business people express a logical in this way . Just ask what your major is . replied ： Mathematical finance , nice ！ This business intelligence BI Most of the business personnel involved in the project are graduate students , Their learning ability is very strong . This is in the follow-up business intelligence BI Tool product training , I feel it .

  Data visualization screen  - Parker data business intelligence BI Visual analysis platform

In the past, I used to do business intelligence for some enterprises on projects BI When training tools and products , Most business personnel are difficult to train . Make reports independently 、 Write some complex functions 、 Look at the data structure , It's really challenging for them . But this business intelligence BI On the project , The training of various departments of this enterprise , I think it's very successful . Business people are easy to learn , At the beginning, if you have any questions, please take the initiative to ask , Later, after proficiency, check the information by yourself 、 Check help documents , Basically, I'll get it done by myself . Sometimes we encounter technical problems , Come and teach me directly , Basically, there is no need to care about the back .

In such business intelligence BI In the project, the simpler our future work is . Monitor every day ETL Data extraction , Have you finished running normally every day . Look at business intelligence BI What new business logic is there in the project 、 Develop the indicators that need to be developed . After development , Maintain new indicators , Write a document to inform you , Basically, the work is finished .

business intelligence BI Project summary

therefore , business intelligence BI In the project IT The best way to divide the boundary with the business is IT Be responsible for the maintenance of underlying data architecture and analysis model , Business personnel make visual analysis reports by themselves . In this business intelligence BI If there is no solid behind the project 、 A highly robust data warehouse cannot support so many users 、 So many report requirements . meanwhile , Business personnel do not have strong data awareness 、 Business ability and learning ability , It is impossible to implement so many reports .

  Data visualization screen  - Parker data business intelligence BI Visual analysis platform

What is good business intelligence BI project , This is a good project , Be steady in the early stage , It looks like a lot of investment , But once the later stage is completed 、 It's landing , It can save at least one business intelligence for enterprises every year BI Resource input of the team , At least millions a year , And the efficiency has been improved .

well , Today's sharing is here , you  Get Have we arrived ？ Focus on big data business intelligence BI, Friends who like our content , Welcome to pay attention to and like support .