expand_matrix {econullnetr}

Expand a summarised interaction matrix

Description

A simple function for converting interaction matrices that are summarised at (typically) species-level to individual-level matrices, ready for use with generate_null_net. This is only applicable to the special (but common) case where one individual = one interaction (e.g. many pollination networks, ant-seed networks). Data can be stored either with consumers as columns and resources as rows or vice versa. Taxon names for each row in the matrix could either be stored as the row names of the matrix or data frame (as used, for example, by the bipartite package), or as a column containing the names in a data frame.

Usage

expand_matrix(X, r.names = rownames(X), MARGIN = 1)

Arguments

X

A matrix or data frame representing the interaction matrix. This should only include the interaction data (i.e. counts of different interactions) and no additional columns for taxon names, covariates, etc

r.names

An optional object of identical length to the number of rows in X listing the taxon names. In many situations these may be the row names of X (the default). Alternatively r.names can be use to specify a column in a data frame containing the names or a separate vector.

MARGIN

Similar to apply, an integer value indicating whether the data are arranged in rows or columns. MARGIN = 1 (the default) indicates that each column relates to one consumer taxon (the format typically used for bipartite networks), whilst MARGIN = 2 indicates that each row is one consumer taxon, with column names being the resources.

Value

A data frame where each row represents the interaction observed between an individual consumer and one resource species. The first column is named Consumer and records which taxon each indidual belongs to. The remaining columns represent the resources: one column for each taxon.

Examples

# Toy example representing a typical bipartite format.
  bp.inter <- matrix(c(1, 2, 2, 0, 5, 3, 3, 0, 2), nrow = 3, byrow = FALSE,
                     dimnames = list(c("A", "B", "C"),
                                     c("sp1", "sp2", "sp3")))
 bp.inter
 expand_matrix(bp.inter)

# Use a simplified version of the Silene data set, pooling data
# across the 11 visits.
  int.summ <- aggregate(Silene[, 3:7], by = list(Silene$Insect), sum)
  colnames(int.summ)[1] <- "taxon"
  expand_matrix(int.summ[, -1], r.names = int.summ$taxon, MARGIN = 2)