Soil_Stats_function.R

##StatisticalAnalysisFunction##
#This function will take a dataframe of soil variables measured at 1cm depths in soil profiles sampled under 2 different land uses.
#It will calculate the mean and standard error of each variable at each depth across replicated plots.
#It will use these to calculate correlations between soil variables and determine if there is an effect of land use.
#In particular I will examine correlations between different phosphorus forms (organic, inorganic, resin, melich, and total), root carbon exudates (phosphatase enzyme, and extractable organic carbon (EOC), as well as the ratio of phosphatase:EOC), and root density.
#The function will return a correlation matrix of the correlations between response variables
# It will also return a dataframe that can be used by subsequent plotting functions. The dataframe will contain columns for mean and SE of each variable and correlations between data variable pairs

#Args
  #dat: argument specifying the dataset of interest. Dataset has previously been imported and correctly formatted according to previous Data_Fill function.

#soil_stats <- function(dat)
{
  #means <- vector of means of each variable at each depth and land use
  #SE <- vector of standard errors of each variable at each depth and land use
  
  #corr <- correlations between variables of interest at each depth and land use. This may be a for loop that will find correlations between multiple variables and store them in multiple vectors to be added to the dataframe.
      #corMat <- the for loop will also generate a correlation matrix
  
  #landuse <- vector of land use names to be repeated and included in dataframe
  #depths <- vector of soil sampling depths to be included in dataframe
  #soilDat<-as.data.frame(cbind(landuse, depths, means, SE, corr))
  
  #return(soilDat, corMat)
}



#For the proposal, I've decided to do a less complex version of soil_stats. This will help me start on the basics while I flesh out the correlation model I need to code
#This function will take a dataframe of soil variables measured at 1cm depths in soil profiles sampled under 2 different land uses.
#It will calculate correlations between soil variables 
#In particular I will examine correlations between different phosphorus forms (organic, inorganic, resin, melich, and total), root carbon exudates (phosphatase enzyme, and extractable organic carbon (EOC), as well as the ratio of phosphatase:EOC), and root density.
#The function will return a correlation matrix of the correlations between response variables
# It will also return a dataframe that can be used by subsequent plotting functions. 

#Args
#soildat: argument specifying the dataset of interest. Dataset has previously been imported and correctly formatted according to previous Data_Fill function.
#var1: first variable of interest, any of the P datasets
#var2: response var to correlate with var1. This will be one of the C datasets.


soil_stats_simple <- function(var1, var2)
{
  correl <- cor.test(x=var1, y = var2)
  return(correl)
}

#Proposed dataframe from data_fill function. This will be used with soil_stats function:
#SoilSummDat <- as.data.frame(4:175)
#colnames(SoilSummDat) <- "Depth"
#SoilSummDat$Pi <- data_fill(PiFra, datReturn = T)
#SoilSummDat$Po <- data_fill(PoFra, datReturn = T)
#SoilSummDat$EOC <- data_fill(EOC, datReturn = T)
#SoilSummDat$AP <- data_fill(AP, datReturn = T)
#SoilSummDat$APEOC <- data_fill(AP.EO)
#soildat <- SoilSummDat[42:172, ]