Skip to main content

Anomaly Detection

Karomia's anomaly detection feature automatically identifies unusual patterns in your emissions data, helping you catch data entry errors, unexpected changes, and areas that need attention.

How it works​

The anomaly detection system analyzes your emissions data and flags entries that deviate significantly from expected patterns. This includes:

  • Sudden spikes — Unusually high emissions in a category or period.
  • Unexpected drops — Significant decreases that may indicate missing data.
  • Trend deviations — Data points that do not follow the established trend.
  • Consistency checks — Values that are inconsistent with related data points.

Viewing anomalies​

Anomalies are displayed on the emissions dashboard:

  • Flagged data points are highlighted in the charts.
  • A dedicated anomaly list shows all detected issues.
  • Each anomaly includes the category, period, expected range, and actual value.

Reviewing and resolving anomalies​

For each flagged anomaly:

  1. Review — Examine the flagged data point and its context.
  2. Investigate — Determine whether the anomaly is a genuine change or a data issue.
  3. Resolve — Either:
    • Confirm the data if the change is legitimate (e.g., new equipment, facility expansion).
    • Correct the data if it was entered incorrectly.
    • Add a note explaining the anomaly for audit purposes.
tip

Regularly reviewing anomalies improves data quality and helps you catch issues early. This is especially important before using emissions data in sustainability reports.

Benefits​

  • Data quality assurance — Catch errors before they affect your reports.
  • Early warning — Identify operational changes that impact emissions.
  • Audit readiness — Demonstrate data validation processes to auditors.

Next steps​