fedivertex 1.1.0

Interface to download and interact with Fedivertex, the Fediverse Graph Dataset

Python API to interact with Fedivertex, the Fediverse Graph Dataset

This Python package provides a simple interface to interact with Fedivertex: https://www.kaggle.com/datasets/marcdamie/fediverse-graph-dataset/data. Our package automatically downloads the dataset from Kaggle and loads graphs in a usable format (i.e., NetworkX).

The Fediverse Graph dataset provides graphs for different decentralized social media. These graphs model the interactions between servers in these decentralized social media. The graph type corresponds to the type of interactions modelled by the graph. Finally, the dataset provides the graphs obtained on different dates, so the users can analyze the evolution of the interactions.

Refer to this repository to discover more about the data acquisition.

Extracting a graph

Three pieces of information are necessary to select a graph in the datatset: the software/social media, the graph type, and the date.

We provide graphs using the NetworkX format.

Example:

    from fedivertex import GraphLoader

    loader = GraphLoader()
    graph = loader.get_graph(software="peertube", graph_type="follow", date="20250324")
    graph = loader.get_graph(software="peertube", graph_type="follow") # Loads the most recent graph

In each graph, we also provide metadata in the attributes of the graph nodes.

Utility functions

Finally, we provide a few utility functions:

    from fedivertex import GraphLoader

    loader = GraphLoader()
    loader.list_all_software()
    loader.list_graph_types("peertube")
    loader.list_available_dates("peertube", "follow")

Cite the dataset

@inproceedings{10.1145/3774904.3792868,
   author = {Damie, Marc and Cyffers, Edwige},
   title = {Fedivertex: a Graph Dataset based on Decentralized Social Media},
   year = {2026},
   isbn = {9798400723070},
   publisher = {Association for Computing Machinery},
   address = {New York, NY, USA},
   url = {https://doi.org/10.1145/3774904.3792868},
   doi = {10.1145/3774904.3792868},
   booktitle = {Proceedings of the ACM Web Conference 2026},
   pages = {8393–8396},
   numpages = {4},
   location = {United Arab Emirates},
   series = {WWW '26}
}