Generalisation of `splnr_plotting.R` function

[KilianBARREIRO]

We want to make the function splnr_plot_cost more generic to handle multi-columns binary and continuous data.

[KilianBARREIRO]

New version :

splnr_plot_cost_4 <- function(df, col_names = NULL, 
                             legendTitle = "Cost", 
                             plotTitle = "",
                             paletteName = "YlGnBu") {


  gg <- ggplot2::ggplot() +
    ggplot2::coord_sf(xlim = sf::st_bbox(df)$xlim, ylim = sf::st_bbox(df)$ylim) +
    ggplot2::labs(subtitle = plotTitle, fill = legendTitle)
  
  # If col_names is NULL, use all columns in the data
  if (is.null(col_names)) {
    stop("Please specify the column(s) of data to display.")
  }
  
  # Replace NA with 0 in selected columns for binary data
  df <- dplyr::mutate_all(df, ~replace_na(.,0))
  
 # Create a new binary column indicating the presence of any non-missing value in selected columns for binary data
  is_binary <- all(sapply(col_names, function(col_name) all(df[[col_name]] %in% c(0, 1))))
  
  if (is_binary) {
    df$UnionColumn <- apply(df[, col_names, drop = FALSE], 1, function(x) as.numeric(any(x == 1, na.rm = TRUE)))
    
    gg <- gg +
      ggplot2::geom_sf(data = df, ggplot2::aes(fill = UnionColumn),
                       colour = "grey80", size = 0.1, show.legend = TRUE) +
      ggplot2::scale_fill_gradient(low = "#c6dbef", high = "#3182bd", limits = c(0, 1))
  } else {
    for (col_name in col_names) {
      gg <- gg +
        ggplot2::geom_sf(data = df, ggplot2::aes(fill = !!rlang::sym(col_name)),
                         colour = "grey80", size = 0.1, show.legend = TRUE) +
        ggplot2::scale_fill_gradientn(
          colors = rev(RColorBrewer::brewer.pal(100, paletteName)),
          limits = NULL
        ) 
    }
  }
  
  return(gg)
}

[jaseeverett]

Following discussion, let's try and make 1 general plotting function that will replace many of the plotting functions in spatialplanr. This generic function should take binary, factorial (many levels such as "epipelagic", "mesopelagic", "bathypelagic" etc:) and continuous data.

This function WILL NOT replace all plotting functions. Let's start with a few first.

• splnr_plot_cost()
• splnr_plot_binFeature()
• splnr_plot_MPAs()
• splnr_plot_featureNo()

Maybe

• maybe splnr_plot_PUs()

[jaseeverett]

@KilianGRIHAULTBARREIRO Can you please provide an update of where you are with this and what functions you intend on working on next?

[KilianBARREIRO]

The function splnr_plot() can replace splnr_plot_cost, splnr_plot_binFeature, splnr_plot_MPAs, and splnr_plot_featureNo.
It works with splnr_plot_PUs , and splnr_plot_solution, but requires display (color/legend) adjustment
It can be generalised to all sf objects (or with a geometry column), whether they contain discrete or continuous values.
I am working on the splnr_plot_costOverlay integration

[jaseeverett]

Ok. I'm not sure what you mean about requiring "display adjustment" but it would be good to get these two (splnr_plot_PUs, and splnr_plot_solution) working properly. If it's something you need help with, let me know.

Each time you have a new function working with splnr_plot, update the vignettes to use the new function so we can all see how to do it. Then we are happy we can delete the old function from the package.

[KilianBARREIRO]

Every function I cited can be replaced by splnr_plot, it's just not the exact same display as the previous one done by the replaced function.
I already replaced those in the vignette, that should be good.

[jaseeverett]

I have added the ability for splnr_plot_PUs to work with splnr_plot

[jaseeverett]

Check if splnr_plot_solution can be removed and replaced with this function instead