R/observation_hist.R

In RoyChristian/R2MCDS: Calling MCDS from R.

#'@export
#'@title Plot a detection histogram of apparent density.
#'
#'
#'@description This function will allow the user to visualise the histogram of observations in function of \code{"distance"} class.
#'@param dataset  A \code{\link{data.frame}} containing observations.
#'@param count Which column in the \code{\link{data.frame}} hold the information for the cluster size.
#'@param dist.class which column in the data.frame hold the information for the \code{"distance"} class.
#'@param keep.class which class of observations should be used for the analysis.
#'@param breaks what are the intervals for the distance analysis.
#'@param color Fill color for the bars.
#'@param ungroup Will plot the Count histogram instead of the group histogram if \code{TRUE}. 
#'@param rescale Set the height of the first bin.
#'@param Type Type of transect data.
#'@details
#' Make a histogram of the observations in function of the distance class
#'@section Author:Christian Roy
#'@examples
#'###Import the data
#'data(alcidae)
#'
#'###Check the naïve detection histogram for alcidae
#'Observation_hist(alcidae,
#'                 count="Count",
#'                 dist.class="Distance",
#'                 keep.class=c("A", "B", "C", "D"),
#'                 breaks=c(0,50,100,200,300),
#'                 color="white")
#'#END


observation_hist <- function(dataset, count, dist.class, keep.class, breaks, color="white",ungroup=F, rescale=1, Type = c("Line", "Point")){
  
  Type <- match.arg(Type)
  ##Keep only the distance class desired
  dataset <- droplevels(dataset[dataset[,dist.class]%in%keep.class,])
  if(ungroup==T){
    ## Make sure to transform groups into individuals
    Observations <- rep(dataset[,dist.class],as.numeric(dataset[,count]))
  }else{
    Observations <- dataset[,dist.class]
  }
  
  levels(Observations) <- sapply(2:length(breaks), function(i){ mean(c(breaks[i-1],breaks[i]))})
  ##Make histogram df
  #browser()
  
  d <- as.numeric(as.character(Observations))
  h <- hist(d, breaks = breaks, plot =FALSE)
  
  if(Type == "Point"){
    h.point <- h$density/(pi*breaks[-1]^2 - pi*breaks[-length(breaks)]^2)
  }
  
  if( !is.na(rescale)==T){
    if(Type == "Line"){
      h$density <- h$density*rescale/h$density[1]       
    }else{
      h$density <- h.point*rescale/h.point[1]        
    }
    
  }
  
  
  r <- data.frame(h[2:4], xmin = head(h$breaks, -1), xmax = h$breaks[-1])
  #Make graph
  ggplot() +
    geom_rect(data=r,aes(xmin = xmin, xmax = xmax, ymin = 0, ymax = density),colour="black", fill=color) +
    scale_y_continuous(limits=c(0,max(h$density))) +
    theme_bw() +
    theme(axis.line = element_line(colour = "black"),
          panel.grid.major = element_blank(),
          panel.grid.minor = element_blank(),
          panel.border = element_blank(),
          panel.background = element_blank()) 
}