Skip to main content

Anomalies & Visualisations on Quantify

Written by Team Enterpret

Overview

Quantify surfaces statistically significant shifts in feedback volume through two chart types:

  • Anomalies — highlights data points that deviate meaningfully from the expected trend.

  • Bar + Trends + Anomalies — overlays the same anomaly markers on top of the bar/trend view so you can see magnitude and anomaly in one glance.

Anomalies are computed using a Z-Score — the number of standard deviations a data point sits from the expected mean for that time window. Default threshold is Z-Score ≥ 1.5.


Use cases

  1. Catch a spike early — detect sudden increases in a theme's volume (e.g. a new bug, outage, or pricing complaint) without manually scanning trends.

  2. Validate a release — confirm that negative feedback didn't spike after a launch, or that positive feedback did.

  3. Monitor competitor or keyword mentions — spot unusual movements in a tracked keyword without setting up separate alerts.

  4. Flag seasonal vs. real shifts — Z-Score normalization accounts for expected variance, so recurring seasonal patterns don't register as anomalies.

  5. Executive reporting — include anomaly-highlighted charts in weekly/monthly reviews so leadership focuses on what changed, not what's stable.


How to discover anomalies in Quantify

  1. Open Quantify from the left navigation.

  2. Build your query: choose a data source, filter, and groupby.

  3. In the chart toolbar, toggle Trends + Anomalies (or select the Anomalies chart type from the chart picker).

  4. Anomaly points appear as marked dots on the chart. Hover any anomaly to see the exact Z-Score and the records driving it.


Configuring the anomaly threshold

  1. In the Quantify chart toolbar, click the gear icon next to the chart type.

  2. Choose Custom under Anomaly Sensitivity.

  3. Use the Z-Score stepper to raise or lower the threshold:

    • Lower Z-Score (e.g. 1.0) — more sensitive; flags smaller deviations.

    • Higher Z-Score (e.g. 2.5) — less sensitive; flags only large deviations.

  4. The chart updates live as you change the threshold.

Tip: Start at the default (1.5). If you're getting too many false positives, raise to 2.0+. If you're missing movements you care about, lower to 1.0–1.25.


Set up anomaly reports

  1. Once your Quantify view is tuned, click Save in the top-right.

  2. Add to Dashboard — select an existing dashboard or create a new one. The anomaly chart becomes a live widget.

  3. Subscribe — click Share → Subscribe to receive a scheduled digest (email or Slack) that includes the anomaly chart and any newly detected anomalies since the last send.

Subscriptions respect the anomaly threshold you configured, so the digest will only call out movements above your chosen Z-Score.


Related articles

Did this answer your question?