Want to come work for seven years, but had never done a systematic summary, the development done, done data, have been in management, who wanted no plan like me, follow the crowd. Recently bitter experience, decided to give himself a positioning (comfort in mind), thought to want to do data analysis for the longest time, knowledge seems to know the most, recognized by the most customers, put itself as a "data management capabilities have analysts "it.
I think the following is a brief record under a data analyst has some of the basic skills, ToB and data analysts ToC companies I have worked for some time, because I learn residue, no chance to enter the business on high-end look tall data analysis of how people are working, only when the following to their summarized.
Data Analyst I contacted roughly be divided into two directions, tend to favor business and technology, in fact, can not be considered in both directions, that is two levels may be more appropriate. Just recruits doing some basic data collection and statistics, this time is biased direction of the business; with the growth of work experience, and slowly exposed to some biased content technologies, such as data mining and machine learning, this time the technical end of the beginning growth, but also to better assist business analysts, data analysts in the end what capabilities it needs it.
My conclusion is that business knowledge, logical thinking, data processing, statistical analysis, analysis tools, data visualization, reporting materials, data mining.
1) business knowledge
No matter which level of data analysts, business knowledge must first have a deeper awareness, an analyst data between business and technology, through the data to tell the business, understand the business only to be able to quickly familiar with the data, and through deepening understanding of the data reverse optimization business.
2) logical thinking
Good logical thinking will not only help to optimize the analysis process, the analysis also plays an important role in preparing the report. Data Analyst have a good logical thinking, analysis reflects the set of ideas, the optimization process of analysis, the conclusions of the report output and for stakeholders to communicate and share.
3) Data processing
From my own experience, the inaugural All the company will require SQL, will SQL, will SQL, basic ORACLE, advanced HIVE, HBASE, as well as more flexible PYTHON.
4) Statistical analysis
As a data analyst, of course, to master the basic method of analysis. If you can only extract some basic data with SQL, or do some reports, it may be in your data development road farther and farther. Basic statistics, the main component factor analysis of variance test to learn.
5) analysis tools
This varies, and now the market has a lot of built-visualization tools, BAT have, on their own Baidu, in addition to the earlier SPSS, SAS, of course, 90% of the work can be used EXCEL.
6) Data Visualization
In my experience speaking, good data visualization does not necessarily convey more information than a normal icon, but will attract stakeholders more attention, in fact, this time data visualization purpose has been achieved, people are willing to look at (ToB particularly evident), the analysis is valuable.
7) briefing material
Do not let the data and conclusions hold in their hearts, SHOW Come out! PPT, WORD, PDF can be, but clearly the center of attention, a clear conclusion, which is the core value of the report.
8) Data Mining
The reason may be due to the work environment, the customer is not satisfied with where I only see the phenomenon, reasons and conclusions, also want to be able to guide them through the steps to develop and improve (in fact, I have written advisory report -.-), and therefore the process of cross Some potential customers with content recognition precision marketing, and similar data product manager. This time some of the data would involve mining, knowledge of machine learning, this does not start speaking, now in full swing the field of artificial intelligence, devoted to the follow-up.
Finally, attach a learning road map data fields.
