Anomaly detection analytic type
Overview
In Insight Analytics, you can use the analytic type Anomaly detection to visualize detected anomalies for a selected attribute.
This analytic type can also be used in Platform.
If you save the configured analytic in Insight Analytics, you can reuse it in dashboards, Smart Views, and reports.
This view lets you:
select an attribute with anomalies
display anomalies in the chart
inspect the anomaly time range
review the anomalies currently visible in the chart
Navigate to: Insight Analytics [1]. ▷ The Insight Analytics page opens.
Elements of the Anomaly detection view
[1]
Insight Analytics
Opens the analytics interface.
[2]
Analytics type
Lets you select the analytic type Anomaly detection.
[3]
Attribute selection
Lets you select an attribute for which anomalies exist.
[4]
Visible anomalies list
Shows the anomalies that are currently visible in the chart.
[5]
Anomaly chart
Shows the selected attribute together with the anomaly visualization.
Open the Anomaly detection analytic type
Open Insight Analytics [1]. ▷ The analytics interface opens.
In the Analytics type selector [2], select Anomaly detection. ▷ The analytics view switches to anomaly visualization.
In the asset tree, select an attribute with anomalies under [3]. ▷ The chart is updated.
◈ In the general analytics interface, the analytics type, location tree, chart options, and time period can be configured from the analytics page .
Understand the anomaly chart
The chart [5] shows the selected attribute in the chosen timeframe.
Anomalies are visualized as follows:
Light blue line: The measured data
Red background: The time range in which the anomaly occurred
Red dot: The detected anomaly point
Green range: The healthy value range predicted by the model
Understand the anomaly tooltip
When you hover over a red anomaly dot in the chart, a tooltip opens.
The tooltip shows detailed information about the selected anomaly.
Financial impact
The estimated financial impact of the anomaly. A negative value means financial loss. A positive value means financial gain.
Time range
The start and end time of the anomaly.
Predicted
The value predicted by the model for healthy behavior.
Actual
The measured value.
Deviation
The difference between the predicted and actual value.
Severity
The severity level of the anomaly.
Validity
The current validation status of the anomaly, for example Confirmed.
Diagnostics
A rule-based diagnostic summary of the anomaly.
AI Explanation
An AI-generated explanation of the likely cause of the anomaly.
Recommended action
An AI-generated recommendation for what to check or do next.
◈ The tooltip helps you review the anomaly directly in the chart without opening the anomaly detail page.
Important note about aggregation and interval
The anomaly time range is highlighted by the red background.
To also see:
the red anomaly dot
and the green predicted healthy range
the selected aggregation mode and interval must match the aggregation at which the anomaly was detected.
If the aggregation or interval does not match, the anomaly can still be marked by its red background, but the anomaly point and the predicted range may not be shown.
For many energy attributes, Sum is typically the useful aggregation function.
Review visible anomalies
The table [4] shows the anomalies that are currently visible in the chart.
The list helps you review the anomalies within the displayed timeframe without leaving the analytics page.
Notes
The Anomaly detection analytic type is part of the Insight Analytics interface .
The analytics interface provides chart options such as chart type, aggregation mode, function, interval, normalization, and unit conversion .
If required, adjust the time period and chart options until the anomaly is shown in the desired level of detail .
Last updated