Mini Project 2: Marvel Universe Social Network

INFO 230 | Spring 2026

Can the structure of a comic book universe reveal sixty years of editorial decisions? This project applies a full network analysis pipeline to the Marvel Universe Social Network from Kaggle (a dataset of 6,439 heroes and 96,104 co-appearances across 12,651 comic issues).

For the full analysis, results, discussion, and interpretations see code/project-workflow.ipynb.

Source: Marvel Universe Social Network — Kaggle

cultural-analytics-project-2/
├── code/
│   ├── project-workflow.ipynb      <- full analysis lives here
│   ├── project-workflow.html
│   └── network_output/
│       ├── marvel_Avengers.graphml
│       ├── marvel_Defenders_and_Mystic_Heroes.graphml
│       ├── marvel_Fantastic_Four.graphml
│       ├── marvel_Spider-Mans_World.graphml
│       ├── marvel_Thor_and_Cosmic_Heroes.graphml
│       └── marvel_X-Men.graphml
├── data/
│   ├── nodes.csv
│   ├── edges.csv
│   └── hero-network.csv
├── figures/
│   ├── interactive_community_0.html
│   ├── interactive_community_1.html
│   ├── interactive_community_8.html
│   └── interactive_community_13.html
├── environment.yaml
├── index.html
├── README.md
└── README.html

1. Clone the repo
2. Create and activate the conda environment:

conda env create -f environment.yaml
conda activate cultural_analytics

3. Launch Jupyter and run code/project-workflow.ipynb top to bottom

All random operations use SEED = 230 so results are fully reproducible.

All dependencies are managed via environment.yaml using the cultural_analytics conda environment. Key packages include scipy, networkx, pyvis, and plotly for network analysis and visualization, alongside pandas, numpy, matplotlib, and scikit-learn for general data processing.

Dataset

• Sanhueza, C. (2017). The Marvel Universe Social Network. Kaggle. https://www.kaggle.com/datasets/csanhueza/the-marvel-universe-social-network

Network Analysis Methods

• Blondel, V. D., Guillaume, J. L., Lambiotte, R., & Lefebvre, E. (2008). Fast unfolding of communities in large networks. Journal of Statistical Mechanics: Theory and Experiment, 2008(10), P10008.
• Newman, M. E. J. (2010). Networks: An Introduction. Oxford University Press.
• Barabási, A. L., & Albert, R. (1999). Emergence of scaling in random networks. Science, 286(5439), 509–512.

Marvel Lore

• DeFalco, T., Sanderson, P., Brevoort, T., Teitelbaum, M., Wallace, D., Darling, A., ... & Cowsill, A. (2019). Marvel Encyclopedia. DK Publishing.
• Gruenwald, M., Grant, S., & Layton, B. (1982). Marvel Super Hero Contest of Champions #1–3. Marvel Comics.

Academic Network Analysis of Marvel

• Alberich, R., Miro-Julia, J., & Rossello, F. (2002). Marvel Universe looks almost like a real social network. arXiv preprint cond-mat/0202174.

Course Materials

• Course Holmes co-occurrence notebook (holmes_cooccurrence_networks.ipynb), INFO 230: Cultural Analytics, Spring 2026.

Claude (Anthropic) was used as an AI assistant throughout this project for help with code structure, debugging, implementation decisions, and writing feedback. All code was read through, understood, and retyped manually. All analytical decisions, interpretations, and conclusions are my own. Claude helped, but the thinking was mine.

By Maddie MacDonald | UC Berkeley Statistics, MA