metalabvienna/co-affiliation-network

Co-Affiliation Analysis (CAA)

Co-Affiliation Analysis (CAA) is a containerised Python toolkit for analysing institutional co-affiliation patterns in scholarly research. It enables researchers to construct co-affiliation networks, fit zero-inflated negative binomial (ZNIB) gravity models, and generate publication-ready visualisations — all in a fully reproducible Docker environment.

This image bundles the complete CAA library together with all CLI entry points required to run the full analysis pipeline without local dependency management.

What this image does

The CAA Docker image allows you to:

• Build co-affiliation networks from author–institution data
• Analyse collaboration dynamics using ZNIB gravity models
• Generate network and model visualisations
• Enrich affiliation edges with travel distance and routing data
• Identify and analyse top-performing research organisations
• Run the entire workflow via command-line tools or configuration files

Key features

• 📦 Fully reproducible Docker environment
• ⚙️ Configuration-driven pipeline (TOML-based)
• 📊 Network construction, modelling, and plotting in one tool
• 🚀 Multiple CLI entry points for modular workflows
• 🧪 Jupyter notebooks available for worked examples (see GitHub)

Source code & documentation

GitHub repository: 👉 https://github.com/christoph-schl/co-affiliation-analyses/

Full documentation, notebooks, and worked examples are available in the repository.

Available CLI entrypoints

Each command runs a specific stage of the CAA processing pipeline. Configuration is provided via mounted volumes.

Create a Default Configuration File

Generate a ready-to-use configuration file for the processing pipeline.

bash
docker run --rm -it \
  --name co-affiliation-network \
  -v "$PWD/config:/app/config" \
  metalabvienna/co-affiliation-network:latest \
  create-default-config

The container creates a config directory (if it does not already exist) and mounts it at /app/config. After the container exits, a config.toml file containing default input paths and parameters will be available.

Create Co-Affiliation Networks

Generate co-affiliation networks and write outputs to $PWD/data/output.

bash
docker run --rm -it \
  --name co-affiliation-network \
  -v "$PWD/config:/app/config" \
  -v "$PWD/data:/app/data" \
  metalabvienna/co-affiliation-network:latest \
  create-network

Fit ZNIB Gravity Models

Generate ZNIB model inputs and fit zero-inflated negative binomial gravity models (if fit_models = true in config.toml).

bash
docker run --rm -it \
  --name co-affiliation-network \
  -v "$PWD/config:/app/config" \
  -v "$PWD/data:/app/data" \
  metalabvienna/co-affiliation-network:latest \
  create-znib-gravity-model

All outputs are written to $PWD/data/output.

Generate Plots

Create all configured network figures, model plots, and supplementary visualisations.

bash
docker run --rm -it \
  --name co-affiliation-network \
  -v "$PWD/config:/app/config" \
  -v "$PWD/data:/app/data" \
  metalabvienna/co-affiliation-network:latest \
  create-plots

Enrich Edges with Travel Routing Data

Attach travel distance and routing metrics to affiliation edges.

bash
docker run --rm -it \
  --name co-affiliation-network \
  -v "$PWD/config:/app/config" \
  -v "$PWD/data:/app/data" \
  metalabvienna/co-affiliation-network:latest \
  enrich-edges

Create Co-Affiliation Networks for Top Performers

Build co-affiliation networks filtered to top-performing research organisations.

bash
docker run --rm -it \
  --name co-affiliation-network \
  -v "$PWD/config:/app/config" \
  -v "$PWD/data:/app/data" \
  metalabvienna/co-affiliation-network:latest \
  create-top-performers-network

Citation

If you use CAA in your research, please cite:

bibtex
@software{caa_2025_v090,
  author    = {Schlager, Christoph},
  title     = {Co-Affiliation Analysis (CAA)},
  year      = {2025},
  version   = {v0.9.0},
  publisher = {Zenodo},
  doi       = {10.5281/zenodo.17972957},
  url       = {https://doi.org/10.5281/zenodo.17972957}
}