Starting from the goal
What is data analysis: It is the process of processing the acquired data with analytical means for a certain problem and discovering business value.
What is the problem: What is the purpose of the business to understand the number of users? We need to help the business to analyze the purpose and decompose the business ideas. Well, this purpose is obviously to verify whether the screening function is valuable, so what indicators to use must be considered, so the activity rate, retention rate, new user needs, user types, etc. are all things to consider.
What is acquisition, figuring out where the data is placed + how to get the data out.
What is data?
The first is database data, including the data of various business systems in the company, system log data, industry data, etc. The second is the data scene, what is called a scene? The most common example is the data analysis index system, such as the "people, goods, and yard" system in the retail industry, in which the three elements of people, goods, and venues can form a data scenario. The establishment method of the index system will be described in later courses. introduce.
What are "means of analysis"?
According to the needs of actual scenarios, it can be divided into four types: descriptive analysis, statistical analysis, mining analysis, and modeling analysis. This is the core content of our later courses, that is, how to analyze in combination with actual business.
For example, descriptive analysis is actually data analysis thinking. The business likes to let you do descriptive analysis, because it does not use algorithms and the like, and the business can understand it, so it can argue with you. Analytical thinking includes basic thinking and goals such as comparison, trend, and segmentation, and mathematical thinking such as deductive reasoning and induction.
This is the real meaning of our data analysis, which is to find abnormalities, reasons, and trends, rather than giving leaders a simple result. Therefore, we should never expect a picture or a model to gain insight into the original appearance of the data; when actually analyzing the business, we should never be limited to one point or a certain model, and we must always follow the essence of data analysis for analysis.