HEOR Software Network Visualization

Overview

This R script visualizes the interdependence of various Health Economics and Outcomes Research (HEOR) software packages in R. The dependencies between packages are represented as a network, where:

• Nodes represent individual R packages.
• Edges represent dependencies, with arrows pointing from the package being depended on to the package using it.
• Node size is determined by degree centrality, making more central nodes appear larger.
• Node color is based on the riskmetric package score (ranging from orange to green, indicating quality and risk level).
• Node shape denotes the type of package (e.g., survival analysis, plotting, utility, etc.).
• Node border color reflects license compliance, with red indicating non-compliant licenses.

Requirements

The script requires the following R packages: - igraph - ggraph - tidygraph - riskmetric - dplyr - tidyr - purrr - visNetwork

Ensure these packages are installed before running the script.

Setup & Usage

1. Clone this repository and navigate to the project directory.
2. Install the required R packages if they are not already installed: r install.packages(c("igraph", "ggraph", "tidygraph", "riskmetric", "dplyr", "tidyr", "purrr", "visNetwork"))
3. Place the list of HEOR-related R package names in data/R_package_names.csv or use the hardcoded example in the script.
4. Run the script in R to generate an interactive visualization of package dependencies.

Key Functionalities

1. Load and Process Package Dependencies

• Retrieves package dependency data from CRAN.
• Filters dependencies to only include those within the HEOR package list.

2. Compute Network Metrics

• Computes degree centrality to determine the prominence of each package.
• Assesses package quality using riskmetric to generate a risk score.

3. Create an Interactive Visualization

• Uses visNetwork to generate an HTML-based interactive network visualization.
• Implements node and edge attributes based on package characteristics.

Output

The script generates an interactive network visualization displaying HEOR R package dependencies. Users can interact with the visualization, highlighting dependencies and selecting specific packages for further inspection.

Applications & similar projects elsewhere

It would be useful to able to understand the dependencies in a given project, ideally at the function level within each package. This is being attempted in the assertHE package. The aim is to extend the function network of a given model to include the external dependencies.

Acknowledgments

This project is developed by Dark Peak Analytics Ltd and integrates insights from various HEOR-related R packages.

For more details, visit Dark Peak Analytics.