Visualize Anomaly Data

Trendz provides three approaches for visualizing anomaly data, suited to different use cases:

Approach	When to use
Anomaly View	Plug-and-play visualization — no configuration required.
Anomaly Fields	Flexible custom views; anomaly data queried directly from the model.
Business Entity Fields	Fine-grained control when anomaly scores are saved to ThingsBoard.

The Anomaly View is a pre-built view type dedicated to anomaly visualization. It provides the same functionality as the Result tab in the anomaly model, but as a standalone view that can be embedded in a ThingsBoard dashboard widget like any other Trendz view.

To create an Anomaly View, click Create View and select Anomalies from the dropdown. The Anomaly View dialog opens — select an existing anomaly model from the list.

Click OK to confirm.

The view opens with the selected model pre-loaded. Click Build report in the top-right corner to load anomaly data for the configured time range.

Available settings:

Setting	Description
Default Anomaly Selection	Which anomaly is highlighted when the view opens: last or max score index.
Hide Anomalies List	Removes the anomaly list table from the view.
Hide Anomalies Chart	Removes the anomaly chart from the view.
Time Range	Time range used for the view.
Filters	Which items are shown in the view.

To add the Anomaly View to a ThingsBoard dashboard, use the Share button in the top-right corner of the view. See Add View on Dashboard for step-by-step instructions.

Anomaly Fields are building blocks for constructing custom anomaly visualizations inside any Trendz view — line chart, bar chart, card, and others. They query anomaly data directly from the model without requiring scores to be saved to ThingsBoard. Multiple Anomaly Fields from different models can be added to the same view.

In any Trendz view, locate the ANOMALY field category in the left panel and drag a field onto the target axis. The following examples use a Line chart, but the same steps apply to any view type.

After dropping the field, a configuration popup opens. Set the Anomaly type, select the Model, choose an Aggregation, and enter a Label. Use MAX to capture peak anomaly intensity at each time point.

To count anomaly occurrences within each time bucket, change Aggregation to COUNT.

Field properties:

Property	Options	Description
Anomaly	SCORE, SCORE_INDEX	Which metric to visualize.
Model	(list of READY models)	The anomaly model to source data from.
Aggregation	AVG, MIN, MAX, SUM, COUNT, LATEST, UNIQ	How values are combined across time or entities.
Label	text	Field name shown in the chart legend.
Chart Type	bar, line, area	Visual representation style.

Aggregation guidance:

SCORE measures anomaly intensity at a point in time:

Aggregation	Suitability
MAX	Recommended — captures peak intensity
LATEST	Recommended — shows the most recent value
COUNT	Recommended — counts anomaly occurrences
SUM, AVG, MIN, UNIQ	Not recommended — lose meaningful signal

SCORE_INDEX measures cumulative anomaly impact over time:

Aggregation	Suitability
SUM	Recommended — shows total anomaly footprint
COUNT	Recommended — counts anomaly occurrences
UNIQ	Recommended
LATEST	Recommended
AVG, MIN	Acceptable but less sensitive

When anomaly score saving is enabled for the model, ThingsBoard stores anomaly scores as regular device telemetry. These can then be added to any Trendz view as Business Entity Fields — the same as any other telemetry key.

The saved anomaly fields appear in the left panel under their business entity, named after the model. They can be added to any view type the same way as any other telemetry field. The example below uses a Pie chart to compare score index totals across devices.

Goal	Use
Quick visualization with no configuration	Anomaly View
Build a flexible custom view across devices	Anomaly Fields
Count anomaly occurrences per time period	Anomaly Fields with COUNT aggregation
Visualize saved scores with precise control	Business Entity Fields