FULLoption: Main function

In TTAinterfaceTrendAnalysis: Temporal Trend Analysis Graphical Interface

Main function

Description

This is the core function of the interface. It receive arguments from the interface (see the function <TTAinterface>) and build regularized time series , perform diagnostics and analyses.

Usage

FULLoption(param, depth=NULL, sal = NULL, site=NULL, rawdata="NO", select="NO", 
resume.reg="NO", test.normality="NO", plotB = "NO", selectBox="ByYears", log.trans="NO", 
plotZ="NO", datashow="NO", help.timestep = "NO", auto.timestep = "NO", time.step = NULL, 
help.aggreg = "NO", auto.aggreg = "NO", aggreg = NULL, mix = "YES", outliers.re = "NO",
na.replace="NO", start = NULL, end = NULL, months = c(1:12), norm = "NO", npsu = 30, 
test.on.remaider = "NO", autocorr = "NO", spectrum="NO", anomaly="NO", a.barplot="NO",
zsmooth="NO", local.trend = "NO", test= "MK", OnOK4=NULL)

Arguments

param	The name of the parameter you want to analyse it must be the name of the column where are your data. Have to be enter like this : "yourparam".
depth	If existing, the depth interval where your data will be analyse. If values are different from depth max and depth min, missing value are exclude Depth column must be name as 'DEPTH'. Enter the value like this : c(a,b). For analysis at one specific depth you can enter c(a,a).
sal	Same thing as for the depth Salinity column must be name as 'S'.
site	Labels of sampling site as they appears in the database Enter the value like this c("S1", "S2").
rawdata	Peform desciptive statistics on raw database, can be "YES" or "NO" (the default).
select	Peform desciptive statistics on selected parameter and site, can be "YES" or "NO" (the default).
resume.reg	Peform desciptive statistics on regularized time series, can be "YES" or "NO" (the default).
test.normality	Perform a Shapiro-Wilk normality test on selected parameter, can be "YES" or "NO" (the default).
plotB	Display a boxplot of rawdata with outliers identified as cirle, can be "YES" or "NO" (the default).
selectBox	Options for plotB: allow to choose between boxplot by years or by months.
log.trans	This option transform your data in log(x+1) prior to perform analysis.
plotZ	Display a plot of the regularized time series, can be "YES" or "NO" (the default).
datashow	Show a table of the regularized data, can be "YES" or "NO" (the default).
help.timestep	Display an advice for time step selection, base on the mean time that separate two successive measurments. Can be "YES" or "NO" (the default).
auto.timestep	Autoexecute the advice without diplay it.
time.step	Choice of the time step for data aggregation during the build of the time series, can be "Fortnight", "Semi-fortnight", "Mensual", "Annual" or "Mono-mensual" for an aggregation of the data of a month of all years (e.g. all January data).
help.aggreg	Display an advice for method of aggregation selection, base on Wilcoxon p.value between rawdata and the different method. Can be "YES" or "NO" (the default).
auto.aggreg	Autoexecute the advice without diplay it.
aggreg	Choice of the method of aggregation during the build of the time series, can be "Mean", "Median", "Max" for maximal value selection or "Quantile" for selection of the quantile 90
mix	Allow to mix the data of all sampling site during analysis. Permanently set to "YES" and removed from the GUI since version 1.5.
outliers.re	Remove the outliers from the rawdata, the outliers list is save in a .csv file. (for outliers visual identification see boxplot section).
na.replace	Replace missing values with median of the corresponding cycle (week, month...) for lags longer than 3 days and linear regression for shorter missed period. Can be "YES" or "NO" (the default).
start	Define the first year of data analysis (by default the first of the database).
end	Define the last year of data analysis (by default the last of the database).
months	Define the months of data analysis (by default the twelve months).
norm	Compute normalised values of nutrients at the salinity npsu for each years, can be "YES" or "NO" (the default).
npsu	Compute normalised values of nutrients at the salinity npsu for each years, 30 by default.
test.on.remaider	Extract the reminders from the data series using the stl package functions to perform statistical analysis.
autocorr	Display the autocorrelation diagramme of the regularized time series using the acf function with arguments : lag.max = ((nrow(TimeSerie))/2), na.action = na.pass. Can be "YES" or "NO" (the default)
spectrum	Display the Fourrier spectrum of the regularized time series using a Smoothed Periodogram (spec.pgram). Can be "YES" or "NO" (the default).
anomaly	Display a color box (function filled.contour) plot by year each time.step (months or weeks) minus the mean of the time.step of all years. Red colors show positive anomalies and blue colors negative anomalies. Can be "YES" or "NO" (the default).
a.barplot	Display an anomaly barplot as a function of the time.step. Red colors show positive anomalies and blue colors negative anomalies. Can be "YES" or "NO" (the default).
zsmooth	Display a detrended plot of the time series using the stl function with arguments s.window="periodic", na.action=na.fail. Can be "YES" or "NO" (the default).
local.trend	Display the interactive cusum plot of the time series (local.trend of the pastecs package) that allow to manually identify the period of change in the tendency using the function identify and perform a Kendall familly test on each idenfified period (see test section). Can be "YES" or "NO" (the default).
test	Perform a test to evaluate the presence and the magnitude of a temporal trend on the time series. Can be "MK" for Seasonal Mann-Kendall test (the default), "SMK" for the same test with detail for each time step, "LOESS" that fit a polynomial surface determined by one or more numerical predictors, using local fitting; a MK is perform on this fitting.
OnOK4	button to call temporal analysis functions

Value

Results are return as .png figures or .txt files Results are also directly readable directly in the right part of the interface.

Savepath can be choose using the option 'Select directory' (see the function selectdirectory more more informations)

Name of saved filed follow the nomenclature : Original.file.name_analysis.name_parameter.txt/.png

or for multiple period analysis (see cusum for more details) : Original.file.name_analysis.name_parameter_starting.year_ending.years.txt.

analysis.names are :

_Boxplot_ for boxplot figure (.png). _Outliers_ for the save of removed outliers (.txt). _TimeSeries_ for the plot of the regularized time series (.png). _Regularised_data_ for the table of regularized time series (.txt). _Autocor_ for the autocorelation diagram (.png). _Spectrum_ for the Fourier spectrum plot (.png) . _ColorPlot_ for the anomaly color.plot (.png). _Anomaly BarPlot_ for the anomaly barplot (.png) _Detrended_ for detrended plot (.png). _Local_Global Trend_ for result of Seasonal Mann Kendall apply to local trend (.txt). _Local_Seasonal Trend_ same as above with detail for each time step (.txt). _Global Trend_ for result of Seasonal Mann Kendall (.txt). _Seasonal Trend_ same as above with detail for each time step (.txt). _LOESSplot_ for loess plot (.png). _NormalNutri_ for analysis of normalized values of nutrients (.png).

See values output of corresponding functions.

Author(s)

David Devreker

See Also

boxplot impute shapiro.test summary acf spectrum filled.contour stl local.trend mannKen seasonTrend seaKen loess

TTAinterfaceTrendAnalysis documentation built on March 31, 2023, 6:12 p.m.