A statisticsPeriodic transformation will compute the configured statistic function. The outputVariable has to be a timeSeries with a defined cycle period. The input periods for a given output time are acquired by repeating the aggregation period for that output time for every cycle. For a given output time the input times in all input periods are used to calculate a result value. A statisticsPeriodic transformation can e.g. be used in climatology to get e.g. the mean temperature in January over the last 100 years. E.g. input series has a temperature value for each day in 100 years and output has a temperature value for each month in the year (this means 12 values in a time series with a cycle of one year).
The StatisticsPeriodic transformation currently only works for a cycle period of one year and only in combination with an output timeStep of type yearlyTimeStep, monthDays, monthlyTimeStep, daysOfMonth or a SimpleEquidistantTimeStep of one day in length. The input timeStep can be anything.
The available statistic functions are:
- count
- countFlags
- kurtosis
- max
- mean
- median
- min
- percentileExceedence
- percentileNonExceedence
- quartile
- rootMeanSquareError
- rsquared
- skewness
- standardDeviation
- sum
- variance
