Calculate percentile values from a time series

Calculates multiple percentile values from a time series, with flexible time aggregation.

Usage

calcPercentile(
  mydata,
  pollutant = "o3",
  avg.time = "month",
  percentile = 50,
  data.thresh = 0,
  start = NA
)

Arguments

mydata: A data frame of data with a date field in the format Date or POSIXct. Must have one variable to apply calculations to.
pollutant: Name of variable to process. Mandatory.
avg.time: Averaging period to use. See timeAverage() for details.
percentile: A vector of percentile values. For example percentile = 50 for median values, percentile = c(5, 50, 95 for multiple percentile values.
data.thresh: Data threshold to apply when aggregating data. See timeAverage() for details.
start: Start date to use - see timeAverage() for details.

Value

Returns a data frame with new columns for each percentile level. New columns are given names like percentile.95 e.g. when percentile = 95 is chosen. See examples below.

Details

This is a utility function to calculate percentiles and is used in, for example, timePlot. Given a data frame with a date field and one other numeric variable, percentiles are calculated.

Author

David Carslaw

Examples

# 95th percentile monthly o3 concentrations
percentiles <- calcPercentile(mydata, pollutant ="o3",
avg.time = "month", percentile = 95)

head(percentiles)
#>         date percentile.95
#> 1 1998-01-01            14
#> 2 1998-02-01             7
#> 3 1998-03-01            15
#> 4 1998-04-01            20
#> 5 1998-05-01            25
#> 6 1998-06-01            15

# 5, 50, 95th percentile monthly o3 concentrations
if (FALSE) { # \dontrun{
percentiles <- calcPercentile(mydata, pollutant ="o3",
avg.time = "month", percentile = c(5, 50, 95))

head(percentiles)
} # }