Anomalies of data


Anomalies of data (flying numbers) in the outputs from the systems usually are usualy visible on the first look, but the time spent searching for causes is many times lengthy. In addition, it may turn out that the anomaly was one-off and caused by a simple error. But time is wasted. The November update of Power BI brings an interesting "preview" function - the search for dependencies of anomalies in the data.

This feature is one of the new "preview" features of Power BI and, depending on the level of sensitivity, will do a lot of work for you. According to the required degree of sensitivity, it analyzes the available data and their parameters and points out possible causes. You choose the parameters yourself or leave it to Power BI and then you just go through the possible causes and decide what to do next.
I tried "Find anomalies" over the data from the service desk and analyzed the anomalies in the quantities of new requests. Interesting what dependencies have turned out ...