01-The beginning of data Analysis-Overview Chapter __ Data Analysis

personal opinions on data analysisAfter doing the data Product manager, has done some simple homework to the data analysis work, now records as follows, hoped can help the data product aspect schoolmate, simultaneously also takes this platform to exchange the study, the improper place, also please treatise.
Data Product Manager ResponsibilitiesData analysis--grasping data dynamics, the value behind the pivot data---------------------------------------------------------------- See the article in detail--the responsibility of the data Product Manager--to be completed.
Data Product IntroductionWhen it comes to data products, you reader will immediately think BI (business intelligence), now the market is also full of a variety of data analysis products, such as tableau such excellent data analysis software, such as Splunk, Saiku, Kylin and other large data analysis products, or enterprise-class data analysis products- Taobao data cube, Google Analytics, Baidu Sinan and so on, are very good data analysis tools, and they are also trial in different fields, will be in the following article on these data products to do analysis (PS: These data products will not be detailed operation guidance, Now there are some tutorials on the software, and these data products also have a better user manual, do not waste time on this. "Tableau product analysis--to be completed" Splunk, Saiku, Kylin data product analysis-pending completion "Taobao data cube, Google Analytics, Baidu Sinan data product analysis-to be completed"
As shown above, now the data products are mainly divided into the following two categories: The free combination of data dimensions and indicators, real-time generation of data reports and visual charts--representative: Tableau; Based on the actual business scenario, regular data is presented as a visual chart for use by the data demand side- Most enterprises build their own bi-class analysis products;
At this point, one might question the repeatability of the two products, "it should be a product." The two should be fused. "Indeed, this is now the ideal of the two states, but due to various restrictions, resulting in two of products in the market coexist, now do the following statement: 1."      Free combination of data dimensions and indicators product representative: Tableau because of this type of product technology difficult to achieve, enterprise-class data volume is also increasing, resulting in enterprises can not do this product, so Tableau should be born, now there are professional to do similar data analysis products companies, such as Yonghong technology; 2. Presentation of the general data in the form of a visual chart representative: Most enterprises set up their own bi-class analysis products why the market has tableau such excellent bi-class analysis software, enterprises have to build their own. At the same time, there is not enough flexibility to do the following analysis: Advantages: (1) Tableau is very expensive, the general enterprise can not pay the high cost of use, (2) data for corporate secrets, do not want to leak; Disadvantage: (1) Self-construction is very labor-intensive, general enterprises do not have a dedicated team to do this thing, to achieve technical difficulties. The better way is to have the data dimension and the index free combination product, in accordance with the actual needs of the summary, and then generate a regular chart for the analyst to use, but because of the above limitations, can only rely on experienced data products and analysts to directly summarize the requirements of the visual analysis products, This is also a lot of company's data developers complain that the development task is heavy, data analyst and data product personnel missing more important reason; (2) The temporary demand still relies on the data developer;
quickly build data analysis productsThen the data products to build the process should pay attention to what problems. Don't visualize it for the sake of visualization, a good data product should not be very good-looking, very cool, but really solve the actual business scene, explore the value behind the data, improve the efficiency of data query, reduce the threshold of use, to achieve product self-taught; see "If" quickly build a data product "
Reprint please indicate the source, thank you ~