Principle:Evidentlyai Evidently Column Statistics Metrics

Knowledge Sources	Evidently Metrics Evidently
Domains	Data_Quality, Statistics, ML_Monitoring
Last Updated	2026-02-14 12:00 GMT

Overview

A collection of column-level statistical metrics for measuring central tendency, distribution shape, and category frequencies.

Description

Column Statistics Metrics provide fine-grained statistical measurements on individual columns:

• MeanValue: Arithmetic mean of a numerical column
• CategoryCount: Count occurrences of specific category values
• InRangeValueCount: Count values within a specified numerical range
• MissingValueCount: Count null/NaN values

These metrics are composable building blocks used both directly in Reports and as sub-components of presets. They enable precise monitoring of data quality properties like value distributions, missing data rates, and category frequencies.

Usage

Use when you need fine-grained column-level statistics beyond what presets provide. Include in Report metrics lists alongside or instead of presets.

Theoretical Basis

Column statistics implement standard descriptive statistics:

# Pseudocode
mean_value = sum(column) / len(column)
in_range_count = count(left <= x <= right for x in column)
category_count = count(x == target for x in column)
missing_count = count(isnan(x) for x in column)

Related Pages

Implemented By

• Implementation:Evidentlyai_Evidently_Column_Statistics_Classes