This article focuses on the most critical stage of the Whatsapp Database entire data analysis process: the data processing and analysis stage. At this stage, I divided it into three parts: data collection, data processing, and data analysis, all of which revolve around "data", Whatsapp Database processing and analyzing massive or messy data, finding pain points and gaining insight into problems. 1. Data collection The data collection here refers to obtaining the data required for analysis, which can generally be obtained from two directions: internal data and external data. 1. Internal data 1) Get it directly .
The premise of direct acquisition is that the company has built a data warehouse and has provided all types of data support for Whatsapp Database decision analysis. This part of the content was also mentioned in the previous article, but here is a little more refined. Direct acquisition means that there are ready-made tables in the database that can directly acquire the required data, without the need for analysts to do complex Whatsapp Database processing on SQL. Companies generally divide data into ods, dwd, and dwb/dws layer data. ods layer: detailed data. The data warehouse does not do any data processing, and directly synchronizes the data to the database intact.
Prepare for the data addition Whatsapp Database of the dw layer. dwd layer: detailed data. The data of this layer has been cleaned on the ods layer, such as removing null values and dirty data. dwb/dws layer: Aggregate data. It mainly summarizes the data of the Whatsapp Database layer slightly, which will involve more business indicator data. For example, according to the detailed data of the ods/dwd layer, indicators such as the repurchase rate of seven days, the comparison data of the same period of the week, and the gross profit rate are calculated for the analyst to directly query.