Dimensions are to be understood in a broad sense. First of all, the time dimension is divided into days, weeks, months and 24 hours. The day dimension is to lengthen the cycle. For example, look at the daily data indicators for one year, look at the trend, look at the daily median, and judge the abnormal time point. Perhaps, when the trend chart is pulled out, it can immediately lock the period of time when the data is abnormal.
As for other time dimensions, it all depends mobile number list on people's different performances in data at different specific times. Weeks represent working days and weekends, and monthly financial settlements, salary payments, etc. may be included, and 24 hours represent working hours, off-duty hours, daytime, nighttime, etc. The time dimension is very special and must be looked at. In addition to the special dimension of time, the rest must be looked at the dimension that is more closely integrated with the business.
In terms of dimension selection, the most common mistake new students make is to drill down without thinking about it, and can't wait to attribute the data item by item. We must try to avoid the explosion of dimensions, and there is no need to drill down into each dimension for data. Originally, it was to explain the reason, and finally it became a bunch of dimensions that exploded.