This brick provides the feature of the rolling window calculation. The concept of rolling window calculation is most primarily used in signal processing and time-series data. Brick takes a window size of k (means k consecutive values at a time) at a time and perform some desired mathematical operation on it.
Сonnect the input dataset to start using the brick, you'll see the dropdown with a list of possible columns for selection, alternatively, you can type the name in manually. You can add and select as many columns as you want by clicking on the '+' button in the brick settings. After that, you need to specify the window size (default is 2) and a function to be applied (sum, mean, min, or max).
- It is highly advised to perform Missing Values Treatment before applying rolling function on data
- Make sure, that chosen columns' types support rolling function (i. e., numerical types)