Resampling¶
Time series data can occur in irregular time intervals. However, to be useful for analytics, the time intervals need to be equispaced.
Resampling recipe¶
The resampling recipe transforms time series data occurring in irregular time intervals into equispaced data. The recipe is also useful for transforming equispaced data from one frequency level to another (for example, minutes to hours).
This recipe resamples all numerical columns (type int or float) and imputes categorical columns (type object or bool) in your data.
Input Data¶
Data that consists of n-dimensional time series in wide or long format.
Parameters¶
Time column¶
Name of the column that contains the timestamps. Note that the timestamp column must have the date type as its meaning, and duplicate timestamps cannot exist for a given time series.
Long format checkbox¶
Indicator that the input data is in the long format. See Long format.
Time series identifiers¶
The names of the columns that contain identifiers for the time series when the input data is in the long format. This parameter is available when you enable the “Long format” checkbox. You can select one or multiple columns.
Time step¶
Number of steps between timestamps of the resampled (output) data, specified as a numerical value.
Unit¶
Unit of the time step used for resampling, specified as one of these values:
Years
Semi-annual
Quarters
Months
Weeks
Business days (Mon-Fri)
Days
Hours
Minutes
Seconds
Milliseconds
Microseconds
Nanoseconds
Interpolate¶
Method used for inferring missing values for timestamps, where the missing values do not begin or end the time series. The available interpolation methods are:
Nearest: nearest value
Previous: previous value
Next: next value
Mean: average value
Linear: linear interpolation
Quadratic: spline interpolation of second order
Cubic: spline interpolation of third order
Constant: replace missing values with a constant
Don’t interpolate (impute null): retrieve missing dates and leave missing values empty
Interpolation methods are based on the scipy implementation.
Extrapolate¶
Method used for prolonging time series that stop earlier than others or start later than others. Extrapolation infers time series values that are located before the first available value or after the last available value. The available extrapolation methods are:
Previous/next: set to previous available value or next available value (if previous values are missing)
Same as interpolation
Don’t extrapolate (impute null): retrieve missing dates and leave missing values empty
Don’t extrapolate (no imputation): don’t extrapolate missing dates
Impute category data¶
Method used to fill in categorical values during interpolation and extrapolation. The available methods are:
Empty : leave the categorical values of the inferred rows empty
Constant: replace missing categorical data with a constant value
Most common: set to the most common value of the time series
Previous/next : set to previous available value or next available value (if previous values are missing)
Previous: set to previous available value
Next: set to next available value
Clip start¶
Number of time steps to remove from the beginning of the time series, specified as a numerical value of the unit parameter.
Output Data¶
Data consisting of equispaced time series, and having the same number of columns as the input data.
Algorithms¶
The resampling recipe upsamples or downsamples time series in your data so that the length of all the time series are aligned. When you specify a given time step (for example, 30 seconds), the recipe will upsample or downsample the time series by an integer multiple of the time step.
The recipe also performs both interpolation (See Interpolate) and extrapolation (See Extrapolate) to infer missing values.