> ## Documentation Index
> Fetch the complete documentation index at: https://alphacastio.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Change Frequency

> Resample a dataset to a different time frequency, interpolating up or aggregating down.

Every dataset has a frequency — daily, weekly, monthly, quarterly or yearly. The **Change Frequency** step converts between them and recalculates the values accordingly.

Moving from a higher to a lower frequency requires an **aggregation** formula, such as average or end of period. Moving from a lower to a higher frequency requires an **interpolation** formula, such as linear interpolation or splines.

<Frame>
  <img src="https://mintcdn.com/alphacastio/ttF_U-V9HDrmrPbo/images/steps/resample-1.png?fit=max&auto=format&n=ttF_U-V9HDrmrPbo&q=85&s=d726bb329358b30526ed0b8eb9a33b3b" alt="Diagram of resampling between time frequencies" width="351" height="111" data-path="images/steps/resample-1.png" />
</Frame>

## How to resample

<Steps>
  <Step title="Create a pipeline and choose the data source" />

  <Step title="Add a Change Sample Frequency step">
    The system detects the input frequency automatically.

    <Frame>
      <img src="https://mintcdn.com/alphacastio/ttF_U-V9HDrmrPbo/images/steps/resample-2.png?fit=max&auto=format&n=ttF_U-V9HDrmrPbo&q=85&s=a455cb0034d57679328e58d9628fa547" alt="Change Sample Frequency step configuration" width="640" height="504" data-path="images/steps/resample-2.png" />
    </Frame>
  </Step>

  <Step title="Select the output frequency">
    This determines whether the step interpolates or downscales.

    **Going up in frequency** (monthly to daily) offers 9 interpolation options: Repeat, Linear, Quadratic, Cubic, Polynomial order 3, Piecewise Polynomial, Spline, Krogh, and Fill with 0.

    **Going down in frequency** (monthly to quarterly) offers 6 aggregations: Average, Sum, Min, Max, Last and First.

    <Frame>
      <img src="https://mintcdn.com/alphacastio/ttF_U-V9HDrmrPbo/images/steps/resample-3.png?fit=max&auto=format&n=ttF_U-V9HDrmrPbo&q=85&s=5ccd60c8453bd13e32a7ac2c056135c4" alt="Available interpolation and aggregation methods" width="555" height="363" data-path="images/steps/resample-3.png" />
    </Frame>
  </Step>
</Steps>

<Warning>
  Interpolating to a higher frequency invents data points that were never observed. It is appropriate for aligning series before a join, but the interpolated values should not be treated as measurements.
</Warning>

## Next steps

<CardGroup cols={2}>
  <Card title="Merge branches" icon="code-merge" href="/steps/merge">
    Join series once they share a frequency.
  </Card>

  <Card title="Apply transform" icon="wand-magic-sparkles" href="/steps/transform">
    Frequency-aware transformations like year-over-year change.
  </Card>
</CardGroup>