Skip to contents

Generate Project Dynamics Data

Source: R/generate_data.R
generate_dynamics_data.Rd

Generates synthetic data to simulate the "Dynamics" of a collaborative project—specifically its Developmental Balance—as defined in the CEnTR*IMPACT framework (Price, 2024).

Usage

generate_dynamics_data(
  seed = Sys.time(),
  exclude = NULL,
  na_prob = 0.05,
  weight_set = c(0.78, 0.84, 0.9, 0.95, 1),
  domain_variance = FALSE
)

Arguments

seed

Integer or POSIXct. The seed for random number generation to ensure reproducibility. Defaults to Sys.time().

exclude

Character vector. A list of Domain names to exclude from the dataset (e.g., c("Contexts")).

na_prob

Numeric. The probability (0 to 1) that a weight value will be replaced with NA. Defaults to 0.05.

weight_set

Numeric vector. The pool of numbers from which weights are sampled. Defaults to c(0.78, 0.84, 0.90, 0.95, 1.00).

domain_variance

Logical. If TRUE, introduces strong bias per Domain to ensure their averages differ significantly. Defaults to FALSE.

Value

A data frame with columns:

  • domain: The framework domain.

  • dimension: The specific dimension within the domain.

  • salience: A randomly assigned importance score (from 0.2 to 1.0).

  • weight: A sampled intensity score based on the weight_set.

Details

Controlling Variance: If domain_variance = TRUE, the function applies Hyper-Polarized Sampling. It forces each Domain to choose a "performance tier" (Low, High, or Mixed) with specific probabilities (Low: 45%, High: 45%, Mixed: 10%).

  • Low Tier: Samples weights only from the bottom 2 values of the provided weight_set.

  • High Tier: Samples weights only from the top 2 values of the provided weight_set.

This maximizes the mathematical spread between domains, making it useful for testing sensitivity or visualizing inequality in the Dynamics scoring.

Schema Structure: The function uses a hardcoded schema representing the 5 core domains of the framework:

  • Contexts (4 dimensions)

  • Partnerships (4 dimensions)

  • Research (5 dimensions)

  • Learning (4 dimensions)

  • Outcomes (4 dimensions)

Examples

# 1. Generate standard data
df_std <- generate_dynamics_data(seed = 123)

# 2. Generate "Unbalanced" data (High Variance)
# This is useful for testing low Balance Scores (Sd)
df_var <- generate_dynamics_data(seed = 123, domain_variance = TRUE)

# 3. Generate data with custom weights and no missing values
df_custom <- generate_dynamics_data(
  na_prob = 0,
  weight_set = c(0.1, 0.5, 0.9)
)