There are many articles on the Internet sharing data indicator systems, but very few articles on data labels. In fact, labels, like metrics, are the right arm of data analysis, and both are equally important. In fact, many people do not analyze in depth because of the lack of application of labels. Explain the system today. A simple example: For example, if we want to introduce Mr. Chen, there are three ways of speaking: Indicators: Mr. Chen is 180cm tall and weighs 200kg; Labels: Mr. Chen is 1.8 meters tall, a big fat man; Tag: Teacher mobile number list Chen, did you hear about the black whirlwind Li Kui? This is the intuitive difference between labels and metrics. Data indicators are accurate descriptions of things using data. Such as height, weight, waist circumference, arm length, these are all data indicators.
Labels are processed based on raw data and carry general descriptions of business meanings. A "big fat man" summarizes height and weight at the same time, and "looks like Li Kui" summarizes features such as facial features, stature, and temperament. Indicators vs Labels. Obviously, in contrast, it is more accurate to describe things with data indicators. But labels are just as important. Because in addition to "accuracy", people have more needs. First, not all features can be described by a single data metric. Common indicators are generally continuous variables (such as height 183cm) or ordinal variables (risk level ABCDE). There are also a large number of features, in the form of categorical variables.
For example, product specifications (50ml bottle), color (red, orange, yellow and green), use (such as: home health care, going out protection...) These product features are generally described in the form of labels, which is also the earliest term for "label". source. Second, labels have business implications. For example, just talking about two indicators: height 183 and weight 200 jins, people don't feel much when they hear it, but once they add a label: height 183 + weight 200, very burly/height 183 + weight 200, big fat man. Do you immediately have a picture in your mind? Finally, tags are easier to use by business. Introducing the object and saying "I'll introduce a little loli to you" is far easier to promote the next action than "I'll introduce a girl with a height of 153 and a weight of 85 to you". That's the beauty of labels. Therefore, the construction of the label system is very important, which can not only enrich the materials for data analysis, but also directly promote the implementation of the analysis results.