Data analysis and data mining are the core technologies and the key to big data applications.
Data analysis is important, but in many cases, we do not know how to do it, and in the face of a large amount of data, we cannot start. To sum up, the difficulties often faced are:
-
The purpose of the analysis is not clear
-
Analysis method is not clear
-
The analysis process is not clear
-
Poor analytical thinking
-
Poor ability to interpret data
1. Don't know what to analyze? (for analysis purposes)
Do not know what to analyze, that is, the purpose of the analysis is not clear.
Students often tell me that the leader has given me a lot of data and wants me to analyze it, but I don't know what to analyze? I don't know what to do other than basic statistical summation.
The purpose of analysis is clear, which is the starting point and the end point of data analysis. All analytical work should begin with a business problem, and the results of the analysis should ultimately fall to the business problem.
If the purpose is not clear, subsequent analysis work cannot be carried out.
2. What is the next step? (Analysis process)
Data analysis is not a single operation, but a complex and complete set of operational processes.
Generally, a complete data analysis includes six steps, and the latter step depends on the former step and is also the depth of the former process.
When there is an analysis purpose, the next step is to collect relevant data around business problems, and preprocess the collected data (cleaning, transformation, extraction, calculation), if you use BI tools such as FineBI to process In this case, the data is extracted first, processed by ETL, and then analyzed in multiple dimensions at the front end, and the analysis results are visualized, and finally a complete analysis report is formed. At this point, a data analysis work is officially completed.
3. Don't know how to analyze? (Analytical method)
The purpose of the analysis is clear and the data is available, but in the face of a large amount of complex data, there is no way to start and do not know how to analyze it. This is due to the lack of understanding of the analysis method by the analyst.
The core task of data analysis is to analyze data. Focusing on business problems, what kind of analysis method to use, what kind of analysis model to use, and what kind of analysis tool to choose, this is the core of data analysis. This is an essential skill for analysts.
For ease of understanding, I divided data analysis into three levels, from low to high, from shallow to deep, namely statistical analysis, basic analysis, and data mining.
一般情况下，企业有80%的工作都只需要掌握统计分析方法就可以了，剩下20%的工作需要更深入的分析及挖掘。当然，更深层次的业务规律及业务模式，需要更高层次的数据分析来解决。比如，市场细分，客户特征提取，等等。
4、 看不明白分析结果？（数据解读）
好不容易分析有结果了，统计有数据了，但是，这些数据及分析结果表示什么意思呢？与我们的业务有什么关系呢？这一步也不知道坑了多少学员。
对数据不敏感，解读数据的能力差，无法将分析结果与业务问题和业务策略关联起来，这是数据应用的最大障碍。
如何来解读数据，解读分析结果，这需要有一定的数据解读方法，也需要分析师要了解相应的业务逻辑。
5、 不知道分析是否全面？（分析思路）
我经常收到一些分析师的抱怨，他们说，基本的分析我都会了，但是，每次提交分析报告给领导以后，领导总是不太满意，说我分析不全面，漏此漏那的。分析不全面，这是由于缺乏分析思路导致的。
如果说，分析方法是从微观从细节来对数据进行分析，那么，分析思路，就是从宏观角度指导如何进行数据分析，比如从哪几个方面来进行完整的数据分析而不会遗漏。
要掌握分析思路，需要分析师懂业务、懂管理、懂营销。比如，如果要分析企业的外部环境，你必须要懂得PEST模型，即要从政策、经济、社会和技术四个方面来进行分析，否则就是不全面的；如果要做竞争分析，你需要懂得SWOT、波特五力，从这几个方面来分析竞争态势，才算完整和系统。
最简单，最实用的是5W2H模型，广泛用于企业营销活动、用户行为分析等专题分析中，即要求分析的从下面7个方面来进行分析，这样可以确保能够将用户购买行为分析完整、系统。
数据分析看起来很简单，但如果没有经过系统的培训，要胜任这项工作也是不容易的。毕竟，数据分析师作为企业主管的智囊，作为主管决策的支撑，其重要性及高要求是不言而喻。