Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.siftstack.com/llms.txt

Use this file to discover all available pages before exploring further.

​
Overview

This guide walks you through the process of configuring Sift Family statistics. Family statistics compute aggregate time series across all included Runs in a Family. Once configured, statistics can be plotted alongside individual Runs in Explore and used as dynamic thresholds in Rules. Common aggregates include the mean signal shape, min/max envelope, and standard deviation band across all Runs in the Family.

​
Create a Family stat

You can create a stat from two places:
  • In the Family creation or edit wizard, proceed to the Statistics step and click Add Statistic.
  • On a Family overview page, click Create Family Stat in the Family Analysis section.
Fill in the following fields:

​
Aggregate Channel Group

Select the primary Channel to aggregate. Sift searches all included Family member Runs for a Channel matching that name and uses it as the input for the aggregation. If some Runs use a different name for the same measurement, click Customize to open Customize Channel Fallbacks. Customize Channel Fallbacks lets you define additional Channel names when the primary name is not found in a given Run. The names are tried sequentially. For example, if older Runs use chamber_pressure and newer Runs use chamber_pres, you can set one as the primary and the other as a fallback so all Runs are included in the aggregate. Customize Channel Fallbacks shows a table of all Family member Runs and the Channel name that was resolved for each Run. Runs with no matching Channel across the primary name and all fallbacks are flagged with a warning. Those Runs are excluded from the aggregate for that statistic.
Calculated Channels are not supported as aggregate Channel inputs. Only raw Channels are available for selection.

​
Aggregation type

In Calculated Channels, choose how to aggregate across Runs at each time step: To compute a sigma band (mean ± N standard deviations) for use in Rules, create separate avg and stdev stats, then reference them together in a rule expression. See Build Rules with dynamic thresholds using Families.

​
Output data rate

In Data rate, set the output sample rate in Hz. The default is 1 Hz. Higher rates retain more temporal detail but require more computation and storage. Choose a rate appropriate for the dynamics of your signal.

​
Data ranges

Data ranges defines the time window within each Run used for aggregation. By default, the full Run (from start to end) is used. A dynamic window is a custom time range anchored to an alignment point rather than the full Run. To define a dynamic window:
  1. Click Add range.
  2. Set the T-0 Alignment: the alignment point that anchors the window.
  3. In Dynamic Window, set the window start and window end as offsets relative to T-0. For example, -10s to +30s captures a 40-second window centered on the alignment event. You can also reference a different alignment point for the start or end boundary. For example, start the window at one event (such as engine ignition) and end it at another (such as engine shutdown). This lets you define windows that span between two meaningful events in the Run.
  4. If the alignment point you need for a boundary doesn’t exist yet, click Back and then Add alignment to define it without losing the the stat editor configuration. A default range covering the full Run is always created. Dynamic windows let you focus the aggregate on a specific phase of the test. This is a key capability when comparing time-delimited events across Runs.
  5. Click Save. Sift computes the stat across all included Runs and returns the result for use in Explore and Rules.

​
How computation works

When you save a Family stat, Sift does the following:
  1. Fetches raw data from each included Run within the configured time window.
  2. Converts each Run’s data to relative time using the alignment configuration.
  3. Buckets data by the output sample rate.
  4. Computes the aggregation across all Runs for each time bucket.
  5. Returns the result for use in Explore and rule evaluation.

​
Preview a stat

While editing a stat, click Preview to see a preview of the computed aggregate plotted in the stat editor. This lets you verify the aggregation looks correct before saving.

​
View existing stats

To view existing stats:
  1. In Sift, click the Families tab.
  2. Click the name of a Family to open its overview page.
  3. Scroll to the Family Analysis section. The stats table lists all configured stats with their name, Channel, aggregation type, and data range.

​
Edit a stat

To edit a stat:
  1. In the Family Analysis section, selection the options for a stat, and click Preview/Edit Stat.
  2. Update any field and click Save.
Sift recomputes the stat with the new configuration and creates a new Family version.

​
Duplicate a stat

Duplicating a stat creates a copy with the same Channel, aggregation type, and data ranges. The duplicate is named Copy of [original name] and is inserted immediately after the original in the stats list. Use a duplicate to create variants of an existing stat. For example, to use different fallback Channels or apply a different output sample rate without reconfiguring from scratch. To duplicate a stat:
  1. In the Family Analysis section, selection the options for a stat, and click Duplicate Family Stat.
  2. The duplicate opens in the stat editor. Rename it and adjust any fields as needed.
  3. Click Save.

​
Delete a stat

Check whether any rules reference this stat before deleting. Dependent rules are listed on the Family overview page under Rule references.
To delete a stat:
  1. In the Family Analysis section, click a stat name to open the editor.
  2. Click Delete.