Zero code | easily realize data warehouse modeling and build Bi Kanban

To do cube analysis and processing , Then it is necessary to build a multidimensional analysis model . Next, we will try to talk about the data analysis model design of Jian Daoyun through examples .

When talking about analysis model, first talk about data processing warehouse and Modeling Technology .

1. About data warehouse

Data warehouse （Data Warehouse,DW） It is a technology born by enterprises to process and analyze all the data collected , The problem to be solved is how to deal with data 、 How to analyze data , Different from database technology, it is born for business operation .

The data warehouse has the following 5 Big feature ：

(1) subject-oriented

The data warehouse loads the data of multiple business systems together through one subject domain , For every theme ( Such as ： user 、 Order 、 Commodities, etc ) Analyze and build , Operational database is built to support various businesses .

(2) Integration

Data warehouse will summarize the data from different databases .

(3) historic

Compared with the operational database , The data of data warehouse is established for enterprise data analysis , Therefore, after the data is loaded, it will generally be retained for a long time , The former is usually kept for several months , The latter may be years or even decades .

(4) Time varying

It means that the data warehouse contains data snapshots from different time periods of its time range , With these data snapshots , Users can summarize it , Generate data analysis reports for each historical stage .

(5) stability

Generally, the data in the data warehouse only performs query operations , There are few deletions and updates . But the data needs to be loaded and refreshed regularly .

2. About data warehouse modeling

There are three main models for data warehouse modeling ：

Star mode

Star mode (Star Schema) Is the most common way to model dimensions , The following figure shows the chart relationship structure of dimension modeling using star pattern ：

 

It can be seen that , The dimensional modeling of star pattern consists of a fact table and multiple dimensional tables , And has the following characteristics ：

• Dimension table is only associated with fact table , There is no relationship between dimension tables ;
• The main code of each dimension table is a single column , And the main code is placed in the fact table , As the outer code connecting the two sides ;
• Take the fact table as the core , The dimension table is star shaped around the core ;

Snowflake mode

Snowflake mode (Snowflake Schema) It's an extension of star mode , Each dimension table can continue to connect multiple sub dimension tables outward . The following figure shows the chart relationship structure of dimension modeling using snowflake pattern ：

The dimension table in star mode is larger than that in snowflake mode , And it doesn't meet the standard design . The snowflake model is equivalent to splitting the star schema large dimension table into small dimension table , Meet the standard design . However, this mode is rarely used in practical applications .

Constellation mode

Constellation mode (Fact Constellations Schema) It's also an extension of the star pattern ：

 

The two dimension modeling methods mentioned above are multi-dimensional tables corresponding to single fact tables , But in many cases, there is more than one fact table in the dimension space , A dimension table can also be used by multiple fact tables . Therefore, most of the actual business analysis is modeled with constellation patterns .

3. Data analysis modeling

This article will use the case database found on the Internet , Analyze the modeling of Jiandao cloud .

Here is the passage power bi Design analysis model

The main indicators we need to pay attention to are ： Order quantity 、 sales 、 Product sales and inventory 、 gross profit 、 Customer credit .

The dimensions that need to be analyzed are ：

1. Time dimension 、 Product dimension 、 Employee dimension 、 Customer dimension .
2. There are geographical dimensions in the dimensions of customers and employees .

Next we will base on power bi The relationship model of , Build this model in Jian Daoyun's data factory .

It can be seen that the whole model is divided into 4 Most of the （4 Group summary ）.

• The first part ： Solve the two indicator data of customer credit limit and payment amount , And all dimension data involved .
• The second part ： Solve the order quantity index data , And all dimension data involved .
• The third part ： Solve sales 、 Product sales 、 The selling cost of the product 3 Indicators , And all dimension data involved .
• The fourth part ： Solve the product inventory index , And all dimension data involved .

Finally, by adding and merging 4 Some dimensions and indicators are merged , As shown in the figure below ：

The front end of each part , Are connected horizontally , According to the relationship between tables （ That is, the primary key of each table 、 Foreign key relationships ）, Left connected 、 Right connected .

for example ：

thus , The cube analysis model of all our tables has been built .

The following is the model data , To build the dashboard BI, An analysis chart showing all indicators and dimensions .

4. The dashboard BI Chart display

（1） Overall data display

（2） Order quantity analysis

（3） Sales analysis

This article only does this 3 Part of the dashboard design , You can also customize it according to your actual needs . thus , The whole analysis model construction and display is over .