summaryPlot can also provide summaries of a
single pollutant across many sites.summaryPlot(mydata, na.len = 24, clip = TRUE, percentile = 0.99,
type = "histogram", pollutant = "nox", period = "years",
breaks = NULL, col.trend = "darkgoldenrod2", col.data = "lightblue",
col.mis = rgb(0.65, 0.04, 0.07), col.hist = "forestgreen", cols = NULL,
date.breaks = 7, auto.text = TRUE, ...)date field and at least one other parameter.na.len contiguous missing vales. The purpose of
setting na.len is for clarity: with long time series it is
difficult to see where individual missing hours are. Furthermore,
sclip = TRUE, will remove the top 1to yield what is often a better display of the overall
distribution of the datapercentile = 0.99 (the default) will remove the top 1
percentile of values i.e. values greater than the 99th percentile
will not be used.type is used to determine whether a histogram
(the default) or a density plot is used to show the distribution
of the data.pollutant is used when there is a field
site and there is more than one site in the data frame.period is either years (the default)
or months. Statistics are calculated depending on the
period chosen.colors() into R to
see the full range of colour names.colors() into R to see the full range of colour
names."greyscale".TRUE (default) or FALSE. If
TRUE titles and axis labels will automatically try and
format pollutant names and units properly e.g. by subscripting
the xlab,
ylab and main), which are all passed to the plot via
quickText to handle routine formattisummaryPlot produces two panels of plots: one showing the
presence/absence of data and the other the distributions. The left
panel shows time series and codes the presence or absence of data
in different colours. By stacking the plots one on top of another
it is easy to compare different pollutants/variables. Overall
statistics are given for each variable: mean, maximum, minimum,
missing hours (also expressed as a percentage), median and the
95th percentile. For each year the data capture rate (expressed as
a percentage of hours in that year) is also given.
The right panel shows either a histogram or a density plot
depending on the choice of type. Density plots avoid the
issue of arbitrary bin sizes that can sometimes provide a
misleading view of the data distribution. Density plots are often
more appropriate, but their effectiveness will depend on the data
in question.
summaryPlot will only show data that are numeric or integer
type. This is useful for checking that data have been imported
properly. For example, if for some reason a column representing
wind speed erroneosly had one or more fields with charcters in,
the whole column would be either character or factor type. The
absence of a wind speed variable in the summaryPlot plot
would therefore indicate a problem with the input data. In this
particular case, the user should go back to the source data and
remove the characters or remove them using R functions.
If there is a field site, which would generally mean there
is more than one site, summaryPlot will provide information
on a single pollutant across all sites, rather than provide
details on all pollutants at a single site. In this case
the user should also provide a name of a pollutant
e.g. pollutant = "nox". If a pollutant is not provided the
first numeric field will automatically be chosen.
It is strongly recommended that the summaryPlot
function is applied to all new imported data sets to ensure the
data are imported as expected.# load example data from package
data(mydata)
# do not clip density plot data
summaryPlot(mydata, clip = FALSE)
# exclude highest 5 \% of data etc.
summaryPlot(mydata, percentile = 0.95)
# show missing data where there are at least 96 contiguous missing
# values (4 days)
summaryPlot(mydata, na.len = 96)
# show data in green
summaryPlot(mydata, col.data = "green")
# show missing data in yellow
summaryPlot(mydata, col.mis = "yellow")
# show density plot line in black
summaryPlot(mydata, col.dens = "black")Run the code above in your browser using DataLab