This function generates a multivariate plot for a data matrix (raster), with an option for filtering the data and to plot using groups. The resterplot graph is a flexible tool useful for different data sources.
eco.rasterplot(x, filter = NULL, condition = NULL, grp = NULL, limits = NULL, title = NULL, z.name = NULL, vertical = TRUE, interactivePlot = TRUE, ...)
| x | Data matrix (raster) |
|---|---|
| filter | Optional data matrix used as filter |
| condition | Condition used to filter data |
| grp | Factor with groups to use in the plot. Default NULL |
| limits | Balues limits used for computing the data gradient for the plot |
| title | Plot title |
| z.name | Name for the legend |
| vertical | Should the populations on the x axis be partitioned? Default TRUE. |
| interactivePlot | Show an interactive plot via plotly? (default: TRUE) |
| ... | additional arguments |
# NOT RUN { data(eco.test) require(ggplot2) # using the ecogen object "eco" to perform a multiple-lsa con <- eco.weight(eco[["XY"]], method = "knearest", k = 4, row.sd = TRUE) test.lsa <- eco.lsa(eco[["P"]], con = con, method = "I", nsim = 99, multi = "matrix") # the plot method for this object based in ggplot2, is a resterplot eco.plotLocal(test.lsa, multi = "ggplot2") # adding a factor test.lsa <- eco.lsa(eco[["P"]], con = con, method = "I", nsim = 99, multi = "matrix", pop = eco[["S"]][,1]) eco.plotLocal(test.lsa, multi = "ggplot2") # The generic rasterplot method requires a data matrix, and, as option, a condition # and a filter matrix. The condition is an expression, containing the word "filter" and # logical elements, e.g., "filter < 50", "filter <50 || filter > 2", etc. ). # Filter is used as a logical matrix (TRUE-FALSE, in relation to the passed condition), # for filtering the data. If a condition is passed but not a filter matrix, the condition # is applied over the data matrix, also using the word "filter". # Internally, the multi.lsa plot uses three fundamental elements. # - a data matrix: in the example, ecoslot.OBS(test.lsa) # - a filter matrix: in the example, ecoslt.PVAL(test.lsa); i.e., # the data matrix will be filtered by P-value using the third element, an expresion. # - an expression: in the example: "filter < 0.05" # by combining the three elements, the multivariate plot can be manually constructed: my.plot <- eco.rasterplot(x= ecoslot.OBS(test.lsa), filter = ecoslot.PVAL(test.lsa), condition = "filter < 0.05") my.plot # add population my.plot <- eco.rasterplot(x= ecoslot.OBS(test.lsa), filter = ecoslot.PVAL(test.lsa), condition = "filter < 0.05", grp = ecoslot.POP(test.lsa)) my.plot # extra manipulation with ggplot2 graphs (ggplot2 commands allowed by rasterplot) my.plot <- eco.rasterplot(x= ecoslot.OBS(test.lsa), filter = ecoslot.PVAL(test.lsa), condition = "filter < 0.05", interactivePlot = FALSE) my.plot ## rotate plot my.plot + coord_flip() ## change design my.plot + theme_grey() # using the data as filter eco.rasterplot(x= ecoslot.OBS(test.lsa), filter = ecoslot.OBS(test.lsa), condition = "filter > 0 & filter < 3") # example of bad syntax (incorrect use of && over matrices) eco.rasterplot(x= ecoslot.OBS(test.lsa), filter = ecoslot.OBS(test.lsa), condition = "filter > 0 && filter < 3") # }