Built with Tutify.io

    AIStudyLabs

    Polars Series rolling_mean() – Usage & Examples

    Polars Series rolling_mean() – Usage & Examples

    In Polars, the rolling_mean() function is used to compute the rolling (or moving) average over a window of values in a Series. It’s especially useful for smoothing fluctuations in time series data or identifying trends over recent observations. The method works by applying a sliding window of a specified size and computing the average of the values within each window as it progresses through the Series.

    In this article, I will explain the Polars Series rolling_mean() method, covering its syntax, parameters, and usage, and demonstrating how it returns a new Series with the calculated rolling (moving) average values.

    Key Points –

    <a id="h-syntax-of-polars-series-rolling-mean-method"></a>

    Syntax of Polars Series rolling_mean() Method

    Let’s know the syntax of the series rolling_mean() method.

    
# Syntax of series rolling_mean()
Series.rolling_mean(
window_size: int,
weights: list[float] | None = None,
*,
min_samples: int | None = None,
center: bool = False,
) → Series[source]

    <a id="h-parameters-of-the-polars-series-rolling-mean"></a>

    Parameters of the Polars series rolling_mean()

    The following are the parameters of the polars Series rolling_mean() method

    <a id="h-return-value"></a>

    Return Value

    This function returns a new Series containing the rolling (moving) mean values.

    <a id="h-usage-of-polars-series-rolling-mean-method"></a>

    Usage of Polars Series rolling_mean() Method

    The rolling_mean() function in Polars computes the average of values within a sliding window of a specified size along a Series. It’s commonly used to smooth data or identify trends in time series or numerical datasets.

    Now, let’s create a Polars series using a list of values and use the rolling_mean() function.

    
import polars as pl

ser = pl.Series("numbers", [2, 4, 6, 8, 10])
print("Original Series:\n", ser)

    Yields below output.

    To calculate a simple rolling mean (moving average) with a window size of 3 on your Polars Series, you can use the rolling_mean() method.

    
# Rolling mean with window size 3
ser2 = ser.rolling_mean(window_size=3)
print("Rolling mean with window size 3:\n", ser2)

    Here,

    <a id="h-rolling-mean-with-window-size-2-centered"></a>

    Rolling Mean with Window Size 2, Centered

    To calculate a centered rolling mean with a window size of 2 in Polars, use the rolling_mean() method with window_size=2 and center=True.

    
# Rolling mean with window size 2, centered
ser2 = ser.rolling_mean(window_size=2, center=True)
print("Centered rolling mean with window size 2:\n", ser2)

# Output:
# Centered rolling mean with window size 2:
# shape: (5,)
# Series: 'numbers' [f64]
# [
# 	null
#	3.0
#	5.0
#	7.0
#	9.0
# ]

    Here,

    <a id="h-rolling-mean-with-custom-weights"></a>

    Rolling Mean with Custom Weights

    To compute a rolling mean with custom weights in Polars, use the rolling_mean() function along with the weights parameter.

    
# Custom weights (for window size 3, for example)
weights = [0.1, 0.3, 0.6]  # Must sum to 1 (not mandatory but typical for weighted averages)

ser2 = ser.rolling_mean(window_size=3, weights=weights)
print("Rolling mean with custom weights:\n", ser2)

# Output:
# Rolling mean with custom weights:
# shape: (5,)
# Series: 'numbers' [f64]
# [
#	null
#	null
#	5.0
#	7.0
#	9.0
# ]

    Here,

    <a id="h-rolling-mean-on-floating-point-series"></a>

    Rolling Mean on Floating Point Series

    You can easily compute a rolling mean on a floating-point Polars Series using the same approach.

    
import polars as pl

# Create a floating-point Series
ser = pl.Series("floats", [2.5, 4.0, 6.7, 8.3, 10.1])

# Calculate rolling mean with window size 3
ser2 = ser.rolling_mean(window_size=3)
print("Rolling mean on floating-point series:\n", ser2)

# Output:
# Rolling mean on floating-point series:
# shape: (5,)
# Series: 'floats' [f64]
# [
#	null
#	null
#	4.4
#	6.333333
#	8.366667
# ]

    Here,

    <a id="h-rolling-mean-on-a-series-with-null-values"></a>

    Rolling Mean on a Series with Null Values

    When performing a rolling mean on a Polars Series with null values, Polars automatically skips the nulls within each window during the mean calculation. However, if all values in a window are null, the result is null.

    
import polars as pl

ser = pl.Series("numbers", [2, None, 6, None, 10])

# Rolling mean on a Series with null values
ser2 = ser.rolling_mean(window_size=3)
print("\nRolling mean with window size 3:\n", ser2)

# Output:
# Rolling mean with window size 3:
# shape: (5,)
# Series: 'numbers' [f64]
# [
#	null
#	null
#	null
#	null
#	null
# ]

    <a id="h-conclusion"></a>

    Conclusion

    In summary, Polars’ rolling_mean() method provides a powerful and flexible way to compute moving averages on Series data. Whether you need basic rolling means, centered windows, or weighted averages, it handles null values gracefully and works seamlessly with different data types.

    Happy Learning!!

    <a id="h-related-articles"></a>

    Related Articles

    <a id="h-references"></a>

    References