openair (version 0.4-10)

cutData: Function to split data in different ways for conditioning

Description

Utility function to split data frames up in various ways for conditioning plots. Users would generally not be expected to call this function directly. Widely used by many openair functions usually through the option type.

Usage

cutData(x, type = "default", hemisphere = "northern", 
                 n.levels = 4, is.axis = FALSE, ...)

cutDaylight(x, local.hour.offset = 0,
                 latitude = 51.522393, longitude = -0.154700, 
                 ...)

Arguments

x

A data frame containing a field date.

type

A string giving the way in which the data frame should be split. Pre-defined values are: "default", "year", "hour", "month", "season", "weekday", "ws", "site", "weekend", "monthyear", "daylight", "gmtbst" or "bstgmt". type can al

hemisphere

Can be "northern" or "southern", used to split data into seasons.

n.levels

Number of quantiles to split numeric data into.

is.axis

A logical (TRUE/FALSE), used to request shortened cut labels for axes.

local.hour.offset, latitude, longitude

Parameters used by cutDaylight to estimate if the measurement was collected during daylight or nighttime hours. local.hour.offset gives the measurement timezone and latitude and longitude giv

...

All additional parameters are passed on to next function(s). For example, with cutData all additional parameters are passed on to cutDaylight allowing direct access to cutDaylight via either cutData

Value

• Returns a data frame with a column cond that is defined by type.

Details

This section give a brief description of each of the define levels of type. Note that all time dependent types require a column date. "default" does not split the data but will describe the levels as a date range in the format "day month year". "year" splits the data by each year. "month" splits the data by month of the year. "hour" splits the data by hour of the day. "monthyear" splits the data by year and month. It differs from month in that a level is defined for each month of the data set. This is useful sometimes to show an ordered sequence of months if the data set starts half way through a year; rather than starting in January. "weekend" splits the data by weekday and weekend. "weekday" splits the data by day of the week - ordered to start Monday. "season" splits data up by season. In the northern hemisphere winter = December, January, February; spring = March, April, May etc. These defintions will change of hemisphere = "southern". "daylight" splits the data relative to estimated sunrise and sunset to give either daylight or nighttime. The cut is made by cutDaylight but more conveniently accessed via cutData, e.g. cutData(mydata, type = "daylight", latitude=my.latitude, longitude=my.longitude). The daylight estimation, which is valid for dates between 1901 and 2099, is made using the measurement location, date, time and astronomical algorithms to estimate the relative positions of the Sun and the measurement location on the Earth's surface, and is based on NOAA methods (http://www.esrl.noaa.gov/gmd/grad/solcalc/). The local.hour.offset is zero if you are working in UTC/GMT, otherwise see http://www.srrb.noaa.gov/highlights/sunrise/timezone.html for local time zones corrections. Measurement location should be set using latitude (+ to North; - to South) and longitude (+ to East; - to West). "gmtbst" or "bstgmt" will split the data by hours that are in GMT i.e. mostly winter months) and hours in British summertime. Each of the two periods will be in local time. The main purpose of this option is to test whether there is a shift in the diurnal profile when GMT and BST hours are compared. This option is particularly useful with the timeVariation function. For example, close to the source of road vehicle emissions, `rush-hour' will tend to occur at the same local time throughout the year e.g. 8 am and 5 pm. Therefore, comparing GMT hours with BST hours will tend to show similar diurnal patterns (at least in the timing of the peaks, if not magnitude) when expressed in local time. By contrast a variable such as wind speed or temperature should show a clear shift when expressed in local time for BST vs. GMT. In essence, this option when used with timeVariation may help determine whether the variation in a pollutant is driven by man-made emissions or natural processes. "wd" splits the data by 8 wind sectors and requires a column wd: "NE", "E", "SE", "S", "SW", "W", "NW", "N". "ws" splits the data by 8 quantiles of wind speed and requires a column ws. "site" splits the data by site and therefore requires a column site.

Examples

## split data by day of the week
mydata <- cutData(mydata, type = "weekday")