Question

How do I interpret the results of an anomaly detection model in Power BI?
After applying anomaly detection in my Power BI report, I need to understand what the results mean for analysis and decision-making. What information is provided alongside detected anomalies, and how can I use tooltip data, expected range indicators, and explanations to interpret the significance of each anomaly accurately?

Answer 1

When you apply anomaly detection in Power BI, the results are visualized directly on your line chart with anomaly markers—typically dots or icons—that highlight data points flagged as outliers. Interpreting these results involves understanding the context behind each anomaly, which Power BI provides through tooltips, expected range shading, and AI-generated explanations.

Each anomaly marker comes with a tooltip that shows:

• The actual value

• The expected value range (based on trend and seasonality)

• The timestamp or time period of occurrence

If the actual value significantly deviates from the expected range, it's marked as an anomaly. The expected range is also shown visually as a shaded band around the trend line, helping you quickly spot whether a data point falls outside the normal variation.

To dig deeper, use the "Explain anomaly" feature (available on right-click or via the Analytics pane), which uses AI to analyze contributing factors—like dimension values (e.g., region, product category)—and presents a narrative breakdown or visuals (like waterfall charts) explaining why that point may have been flagged. This helps stakeholders not only see when something went wrong but also why it might have happened, making it easier to take informed action.