Skip to contents

Cast Directed Dyadic Data into Array Format

cast_array.Rd

Transforms long-format dyadic data (edge list) into a 3D array suitable for SIR model analysis. Handles both dyadic and monadic variables with proper placement in the network structure.

Usage

cast_array(dyad_data, var, monadic = FALSE, row = FALSE)

Arguments

dyad_data

Data frame in long format with columns i, j, t, and the value variable specified by var parameter.

var

Character string naming the column containing the values to be placed in the array.

monadic

Logical indicating whether the variable is monadic (node-level) rather than dyadic (edge-level). Default is FALSE.

row

Logical, only used when monadic=TRUE. If TRUE, uses sender attributes; if FALSE, uses receiver attributes. Default is FALSE.

Value

Three-dimensional array (m x m x T) where:

  • First dimension: Sender nodes (i)

  • Second dimension: Receiver nodes (j)

  • Third dimension: Time periods (t)

Missing edges are filled with zeros.

Details

This function converts network data from "long" format (one row per edge/time) to "array" format (3D array indexed by sender, receiver, time).

Expected Input Format: Data frame with columns:

  • i: Sender node identifier

  • j: Receiver node identifier

  • t: Time period

  • var: The variable value for this edge

Monadic Variables: When monadic=TRUE, the function treats the variable as node-level rather than edge-level. The values are placed on the diagonal of each time slice:

  • row=TRUE: Uses sender (i) attributes

  • row=FALSE: Uses receiver (j) attributes

Missing Data: Missing edges in the input are filled with zeros in the output array. This assumes that absence of an edge means zero interaction.

Examples

if (FALSE) { # \dontrun{
# Create example dyadic data
dyad_data <- data.frame(
  i = c(1,1,2,2,3,3),
  j = c(2,3,1,3,1,2),
  t = c(1,1,1,1,1,1),
  trade = c(100,150,80,120,90,110)
)

# Convert to array
trade_array <- cast_array(dyad_data, "trade")
dim(trade_array)  # 3 x 3 x 1

# Monadic example (node GDP)
node_data <- data.frame(
  i = rep(1:3, each=3),
  j = rep(1:3, 3),
  t = 1,
  gdp = rep(c(1000,2000,1500), each=3)
)

gdp_array <- cast_array(node_data, "gdp", monadic=TRUE, row=TRUE)
diag(gdp_array[,,1])  # Shows node GDPs on diagonal
} # }