Chapter 5: Detecting Anomalies in Your Data Using Power BI
Anomaly detection is used when unexpected and rare events need to be identified. Some anomalies are clear outliers and show up as easily recognizable spikes in data. However, some anomalies are more subtle than that and require machine learning (ML) to be detected.
What an anomaly represents depends on the situation. Most commonly, anomaly detection is used for predictive maintenance, when monitoring—for example—an engine. The temperature of an engine can be measured and visualized and is expected to fluctuate. However, whenever the temperature increases or decreases significantly and unexpectedly, it is a cause for concern and a reason to further investigate.
Power BI has an out-of-the-box anomaly detection feature that can be used to detect unexpected events in time-series data. Having this feature makes it very easy for users to find out which data points don't fit in the normal trend of the data...
