Back to datasets
Dataset assetOpen Source CommunitySocial Media AnalysisUser Behavior
Twitter/TwitterFollowGraph
TwitterFollowGraph is a bipartite directed graph comprising user (consumer) nodes and author (producer) nodes, where edges represent a user's "follow" interaction with an author. Each edge is assigned to a predefined time chunk, denoted by a consecutive ordinal that respects the temporal order of interactions. TwitterFollowGraph contains a total of 261 million edges and 15.5 million vertices, with a maximum degree of 900 000 and a minimum degree of 5. The data format is shown in the table below: | user_index | author_index | time_chunk |
Source
hugging_face
Created
Nov 28, 2025
Updated
Oct 31, 2022
Signals
111 views
Availability
Linked source ready
Overview
Dataset description and usage context
Dataset Overview
Dataset Name
- TwitterFollowGraph
Dataset Description
- TwitterFollowGraph is a bipartite directed graph that includes user (consumer) nodes and author (producer) nodes, where edges represent the action of a user "following" an author. Each edge is allocated to a predefined time slot, expressed as an ordinal; the ordinals are consecutive and follow the chronological order.
Dataset Scale
- The graph contains 261 million edges and 15.5 million nodes.
- Maximum degree: 900 000, minimum degree: 5.
Data Format
| Field | Description |
|---|---|
| user_index | User index |
| author_index | Author index |
| time_chunk | Time‑chunk ordinal |
License
- This dataset is released under the Creative Commons Attribution 4.0 International License.
Citation Information
@article{el2022knn,
title={kNN-Embed: Locally Smoothed Embedding Mixtures For Multi-interest Candidate Retrieval},
author={El-Kishky, Ahmed and Markovich, Thomas and Leung, Kenny and Portman, Frank and Haghighi, Aria and Xiao, Ying},
journal={arXiv preprint arXiv:2205.06205},
year={2022}
}
Need downstream help?
Pair the dataset with AI analysis and content workflows.
Once the source passes your review, move straight into summarization, transformation, report drafting, or presentation generation with the JuheAI toolchain.