In recent years, data analysis has become more and more like a general skill development trend. From production, R&D, marketing, sales to operations, there will be more or less demand for data analysis.
Regarding data analysis , there are many cases of analysis reports on the Internet, but after careful reading, many of them lack dialectics, are not rigorous in logic, or have only a shallow analysis. Coincidentally, I recently read " The Dao and Techniques of Big Data Analysis", which is a very complete set of theoretical books. Based on my years of data experience, I have accumulated some experience and want to share it with you.
As a materialist, I always love to talk about methodology. Once "the countryside surrounds the city, and the armed forces seize power" as a synonym for the methodology, it is always on the lips. So what is the methodology of data analysis? It is "first the way, then the surgery, and the way to control the surgery", that is, first understand the core principles of data analysis, and then master some key techniques of data analysis.
What is data analysis? Data analysis refers to the process of analyzing a large amount of data collected with appropriate statistical analysis methods, and conducting detailed research and generalization of the data in order to extract useful information and form conclusions. In fact, it is simple to understand that data analysis itself is the work of "data" and "analysis". On the one hand, it collects, processes, and organizes data, and on the other hand, it analyzes the data in combination with the data background, and extracts conclusions that are helpful to the business.
The output of data analysis is usually a data analysis report . The common ones are PPT and PDF files with graphs and tables, and there are also some web-based graphing systems (if you are interested, you can search for "data analysis" in the app store, and there are various presentation systems that you can experience). For data analysis reports, similar to argumentative papers, the analysis is the argument and the data is the argument.
What is the use of data analysis? This is often asked, especially when a fancy report is published on the Internet, some people say that the report is useless, and the data analysis is useless.
The great value of data analysis to enterprises is reflected in the early stage of business development (exploration period) or the staged improvement period (subversion period). When exploring and transforming business, companies need data analysis to identify problems, opportunities and solutions in the business. The biggest cost of an enterprise is the decision-making cost, and data analysis is the key to improving the decision-making ability of the enterprise; when the business model is relatively mature, the enterprise needs data modeling to improve business efficiency and reduce operating costs.
Then, what are the common and typical application scenarios of data analysis? The author's experience is mainly three: grasping business status, analyzing business potential, and evaluating business progress.
Scenario - Mastering the business status : We need to monitor, interpret and analyze the core indicators to grasp the business operation status. For example, if the sales volume fluctuates abnormally on a certain day, data analysis is required to locate the cause. (The sales volume of product A increased abnormally by 10% this week. What is the reason? How to analyze it?) At the end of the month and the end of the quarter, make a summary of key business indicators, and do business development trend analysis. (What are the key data indicators for this quarter? How is the progress of various businesses? What are the positive and negative factors, and how much will they affect? How will the business develop in the next month and next quarter?) For example, the data platform built by finereport is as follows. Monitored business data. Displayed according to the business analysis rules formulated, you can clearly see which link went wrong.
Sales Analysis
Sales team analysis
Sales Metric Tracking
Scenario - Analyzing Business Potential : What are the current major problems with the product? Where is the next development potential? Mining the cause of the problem from the data and put forward countermeasures, indicating the next improvement direction of the product. For example, commodity B has lost 1,000 members in the past 3 months. What is the reason (analyze the reasons)? How to reduce the loss of members (find countermeasures)?
Scenario - Assessing business progress : How much performance improvement has the newly launched product strategy or newly promoted operational activities brought? What is the reach and impact of the project? What kind of problems exist, how to further optimize and so on. For example, for some customers, a new promotion strategy is designed. After the implementation this month, how to evaluate the performance improvement brought by the promotion strategy? Could an increase in target customer base over last month's purchases be the result of a promotional strategy?
Since data analysis has so many application scenarios and use values, how can data analysis be done well? It can only be said that it is too difficult. It requires both the mastery of tools and skills, and the business experience that can see things clearly. But in comparison, I personally think the latter is more important, just like if you only know the methodology, but don’t have an understanding of the business and how to implement the routine, the data analysis will only be superficial, and it will neither dig out the cause of the problem nor provide solution.
To do a good job in data analysis, regardless of technology, first recognize the following four keys.
1. Business research: understanding the business is the foundation, otherwise the analysis is a tree without roots, or even personal obscenity.
2. Innovative thinking: Broad knowledge and positive thinking are the source of analysis ideas. Innovative thinking in data analysis is essentially to analyze from more ideas to find the most reasonable ideas.
3. Logical reasoning: make correct attribution analysis and judgment on data indicators.
4. Feasible suggestions: Generate practical and effective improvement suggestions and implementation plans for the business.
"Business research" is the starting point of data analysis and the basis for obtaining analysis ideas, but it requires "innovative thinking" with both depth and breadth to obtain more unique analysis ideas. Analysis ideas can also be considered from the perspective of statistical data. After data statistics are completed, "logical reasoning" is required to ensure the correctness of judgments from data to conclusions. Finally, use "feasible suggestions" to ensure the implementation of the analysis conclusions and produce quantifiable performance. This is the process of data analysis coming from the business and back to the business.
This article is the first part of the author's "Big Data Analysis Dao and Technique" reading notes, which is briefly described. Next week, based on my own experience, I will interpret and analyze business research, innovative thinking, logical reasoning, and feasible suggestions one by one.
This article first published CSDN: http://blog.csdn.net/yuanziok/article/details/74908830
Text| Captain of FanRuan Data Application Research Institute