R/DistFitLong.R

In huntercole25/EcoCountHelper: An assistant for comparing and interpretting models with count data

#' @rdname DistFitLong
#' 
#' @export

DistFitLong <- function(Splitters, Data, CountCol, GroupCol,
                        GroupList, ThemeBlack = T){
  .data <- NULL
  
  Plotter <- function(GroupName){
    FullData <- get(deparse(substitute(Data)))
    Count <- FullData[[CountCol]][FullData[[GroupCol]] == GroupName]
    
    Params <- list()
    for(i in Splitters){
      Params[[i]] <- FullData[[i]][FullData[[GroupCol]] == GroupName]
    }
    
    VarData <- stats::aggregate(Count, Params, FUN = stats::var)[[length(Splitters)+1]]
    MeanData <- stats::aggregate(Count, Params, FUN = mean)[[length(Splitters)+1]]
    
    MeanVarData <- as.data.frame(cbind(VarData, MeanData))
    
    if(ThemeBlack == T){pocol = "grey80"}else{pocol = "black"}
    
    TmpPlot <- ggplot2::ggplot(data = MeanVarData, ggplot2::aes(x = .data[["MeanData"]], y = .data[["VarData"]])) +
      ggplot2::geom_point(color = pocol) +
      ggplot2::geom_smooth(method = "lm", formula = y~x-1, ggplot2::aes(color = "NB1"), lty = 1) + #NB1
      ggplot2::geom_smooth(method = "lm", formula = y ~ I(x^2) + offset(x) - 1, ggplot2::aes(color = "NB2"), lty = 1) + #NB2
      ggplot2::geom_abline(ggplot2::aes(intercept = 0, slope = 1, color = "Poisson"), lty = 2) + #Poisson
      ggplot2::scale_color_manual(name = "Families", values = c("green", "purple", "red")) +
      ggplot2::labs(title = GroupName, x = "Mean", y = "Variance") +
      ggplot2::theme_light() +
      ggplot2::theme(plot.title = ggplot2::element_text(hjust = 0.5, face = "bold"))
    
    if(ThemeBlack == T){
      TmpPlot <- TmpPlot + theme_nocturnal()
    }
    
    assign(paste0(GroupName, "DistPlot"), TmpPlot, pos = .GlobalEnv)
  }
  lapply(GroupList, FUN = Plotter)
}