ewma

statistics

Calculates Exponentially Weighted Moving Average (EWMA) for smoothing time series data

Syntax

ewma(column, alpha, min_periods?)

Parameters

column (string)

Column name to calculate EWMA on

alpha (number)

Smoothing factor between 0 and 1 (higher = more weight on recent values)

min_periods (number) optional

Minimum number of observations required (defaults to 1)

Returns

dataframe or array

DataFrame or Array with additional {column}_ewma field/column

Examples

Calculate EWMA with alpha = 0.3

Input:

.price | ewma(0.3)

Output:

[12.5, 14.3, 15.8, 17.2, ...]

EWMA with minimum 5 periods requirement

Input:

.value | ewma(0.2, 5)

Output:

[null, null, null, null, 42.3, 43.1, ...]

Track smoothed trends

Input:

{ original: .sales, smoothed: (.sales | ewma(0.4)) }

Output:

{"original": [100, 120, 110], "smoothed": [100, 112, 111.2]}

The ewma() function calculates Exponentially Weighted Moving Average (EWMA) for smoothing time series data. The formula is: EWMA(t) = α × Value(t) + (1 - α) × EWMA(t-1).

Usage

Use ewma() for smoothing noisy time series data and identifying trends. Common alpha values: 0.1-0.2 for heavy smoothing, 0.3-0.5 for moderate smoothing, 0.6-0.8 for light smoothing. To convert span to alpha: alpha = 2 / (span + 1).

Related Functions

rolling_std mean median