metadata
license: cc-by-4.0
language:
- en
- es
- fr
- pt
size_categories:
- 1B<n<10B
Discord Unveiled Dataset
📄 CC-BY-4.0 Licensed
Dataset Overview
this repo contains the dataset from:
https://arxiv.org/abs/2502.00627
Discord Unveiled: A Comprehensive Dataset of Public Communication (2015-2024), the most extensive Discord public server's data to date. The dataset comprises over 2.05 billion messages from 4.74 million users across 3,167 public servers, representing approximately 10% of servers listed in Discord's Discovery feature. Spanning from Discord's launch in 2015 to the end of 2024, it offers a robust temporal and thematic framework for analyzing decentralized moderation, community governance, information dissemination, and social dynamics. Data was collected through Discord's public API, adhering to ethical guidelines and privacy standards via anonymization techniques. Organized into structured JSON files, the dataset facilitates seamless integration with computational social science methodologies.
2.05B
Messages
4.74M
Users
3,167
Servers
2015-2024
Time Span
Dataset Size & Information
2TB. I have now personally uncompressed the entire thing >:3, using btrfs with zst compression the entire dataset ends up only 175GB on disk! You can also use a command like pzstd -d -c "{{path_to_dataset}}" | tar -xvf - -C "{{destination_dir}}" "./{{guild_id}}.json" to only extract the one file you want to disk. Though it will still need to read the prior possible 1.9TB before extracting the file you want.
Decompression Instructions
To decompress the dataset, use:
Terminal Command
zstd -d -c {file_name}.zst | tar -xvf - -C {output_path}
Server Themes & Keywords
Top 40 most common keywords in Discord server descriptions and their percentages:
What's Actually in Here :P
Important Notice
Data Processing Required: the data will probably need *extensive* cleaning and more thorough PII removal to get to a usuable state :3
Data Collection Disclaimer
All data was collected from public Discord servers using the discovery feature and that collection followed Discord's ToS.