What do indicators and labels do

One 、 Preface

A friend asked me , What is an indicator , What is a label , What is the most essential difference ？ How to identify ？ Don't ask me , I think I'm clear , It seems very clear , But when asked , I don't think I'm clear anymore , So I learned it again , Share your learning notes , Hope to help and inspire you .

 

Two 、 Indicator understanding

1. Index is a concept that describes the overall comprehensive quantitative characteristics , All indicators can be expressed in numerical values , A complete statistical index , Be sure to talk about time 、 place 、 Range （ Baidu ）;

2. The evaluation of indicators is easy to quantify , There are usually certain standards and scales ;

3. The indicator is productive thinking 、 Disassembling thinking , The stress is to break up the whole into parts , Break things down and describe them from multiple perspectives , Get a lot of indicators ;

4. The best application of indicators is monitoring 、 analysis 、 Evaluation and modeling ;

5. Indicators are business management oriented , You need to plan ahead , Many application scenarios , strategic target 、 market positioning 、 Business monitoring 、 Performance appraisal 、 Task breakdown 、 Data analysis 、 Data modeling 、BI Application etc. .

3、 ... and 、 Label understanding

1.  Tags are attributes of objects , Granularity to field level “ label ” It refers to the cleaning and processing of raw data , Data resources that can be used by the business and generate value , Generally, it needs to be structured to field granularity , Ensure service-oriented use .（ Label category system ）

2. Labels are synthetic thinking 、 Convergent thinking , It's about breaking up into whole , Integrate multiple scattered indicators according to certain principles , Get a general result ;

3. Tags are often called attributes 、 features 、 indicators 、 Parameters, etc. ;

4. The indicator is semi-finished products , Labels are finished products , Labels are the result of further commercialization of indicators ;

5. The tag faces the data application side , The answer is “ How to use data ”“ What's the value of data ” The problem of ;

6. Tags are resources , It's assets , Can be priced 、 Can be sold 、 A tradable data product ;

8. Labels are application-oriented , Change with business needs , Add at any time ;

9. The best application of labels is labeling 、 Characterization 、 Classification and feature extraction ;

10. Labels are mainly used in customer clustering 、 portrait 、 Touch guest 、 Receive visitors 、 Sticky guest 、 Follow up 、 Data modeling 、 Data visualization, etc ;

11. The evaluation of labels is generally related to the user's feelings 、 The results of the application have a strong correlation , Different people 、 Different application scenarios , The effect of labels may be quite different .

Four 、 Label layering

1.  Understand the root directory 、 Label categories 、 label 、 The difference and connection between the four tag values , The label system is relatively clear . The following is the thinking of insurance asset level , You can think about the architecture design of the data center .

2.  The root directory points to the object to which the label belongs ： The root directory is often a vague 、 broad 、 Simple nouns or gerunds , For example, users 、 Buyers 、 The hotel 、 Browse （ Record ）、 transaction （ Record ）、 Application for repair （ Record ）. Think in terms of data , Everything in the world can be classified as human 、 matter 、 Three types of objects are related , So a word used to point to an object （ Nouns point to people 、 matter , Gerund pointing relationship ） Should not be labels , It is often the label root directory . At the physical level of data, it is often mapped to the primary key in a large and wide table , The information in this large and wide table is a detailed description and data record of the primary key object ： The columns of the wide table are mapped to labels , The row record of the wide table corresponds to the specific attribute value record of the specific object on each tag attribute .

3. Category is the classification of labels ： Customer labels can be classified as basic information 、 Location 、 Social relationships, etc , These classification names are also category names . Categories are often composed of nouns . A category and its classified tags can correspond to a specific table at the data physical level , for example “ Customer ” Object's 【 essential information 】 Under the heading , Yes “ Gender ”“ Age ”“ Native place ” Wait for multiple tags , Generally, it corresponds to a basic customer information table in the customer database , There will be “ Gender ”“ Age ”“ Native place ” And so on .

4.  Tags are attributes of objects , Granularity to field level ：“ Name of customer ”“ Customer phone ”“ Customer's residential address ” The attribute of equal field granularity is “ Customer ” The label of the object . Labels are often composed of two nouns , The former noun modifies the latter noun as an object attribute . Labels generally correspond to a field in a data table in a database .

 

5.  The tag value is the specific value of the object attribute ： for example 【 Xiao Ming 】【 Xiaohong 】 yes “ Name of customer ” The tag value of the tag ,【 male 】【 Woman 】 yes “ Gender ” The tag value of the tag . Tag values are often adjectives 、 Nouns or numbers , Generally, it corresponds to the value of a field in a data table in the database . The value type of tag value can be numeric 、 The text type 、 Date type 、Key-value type , But it is mainly numerical . Numerical type can be divided into enumerable discrete values and non enumerable continuous values .

5、 ... and 、 Classification of labels

The classification of tags is for applications , You can add as needed .

1.  According to the variability of labels, they are divided into static labels and dynamic labels ;

2.  According to the label reference and evaluation indicators , It can be divided into qualitative label and quantitative label ;

3.  The way of grading and layering according to label assets , It can be divided into primary labels 、 Secondary label 、 Level 3 labels, etc , The label of each level is equivalent to the facet of a business dimension , accord with MECE principle .

5.  It is divided into ： Fact labels 、 Rule tags and model tags . Fact labels are usually realistic , It has a high coincidence with the index ,

6.  Like gender , Age etc. ; Rule tags usually have some simple rules to control , Only when a certain rule is met will the corresponding label be generated ; Model labels generally need to be generated by some machine learning algorithms .

6、 ... and 、 Conclusion

Tags are attributes of objects , Generally, to the field granularity , Data oriented application side , It's resources , It's assets , Can be priced 、 Can be sold 、 A tradable data product , Include properties 、 features 、 indicators 、 Parameters, etc. ;

Indicators are quantifiable , Fields represented by numbers , Business oriented management , You need to plan ahead , The application we are good at is monitoring 、 analysis 、 Evaluation and modeling ;