Should data analysis indicators be determined by data sources or should data indicators be established by data analysis requirements, and data sources should be developed to obtain corresponding data? Theoretically it is the latter, which cannot be limited by the scope of the existing data. Data requirements should be generated based on business objectives and data analysis purposes, and then the requirements should be compared to those data required for support.
However, the objective conditions of data mobile number list acquisition should also be considered. In actual situations, due to the limitation of current technology, some data cannot be collected in a short period of time, although the data is mainly used for analysis. At this time, it may be necessary to consider the idea of ​​​​transforming data analysis, and choose a detour appropriately.
By the way, when using data, it is necessary to verify the accuracy of the data. Sometimes, if the data source to be used has been proven to be reliable enough, verified or matured, we can skip the verification of the data source or slightly verify the data and use the data. However, if the acquisition of data is unstable (common in data of a monitoring nature, involving data collection, data storage, etc., there is a lot to say here), clarify the source of data errors and the acceptability of errors.
For example, to obtain the number of servers in the computer room, the data is relatively reliable, and the interface is widely used. However, when it comes to bandwidth, monitoring, traffic, etc., it is often necessary to mine the data source side, including how the data is collected and storage methods.
For example, in the data of bandwidth utilization, the technical conditions at that time are: one data point will be collected every 5 minutes, but after more than seven days, only one data point will be kept per day. In this way, it is impossible to achieve a daily accurate trend data chart of bandwidth utilization. This is where data affects functionality.