Pith. sign in

REVIEW 4 major objections 8 minor 3 cited by

Discord Unveiled: A Comprehensive Dataset of Public Communication (2015-2024)

T0 review · 4 major / 8 minor · reviewed 2026-08-09 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read This paper claims to release the largest public corpus of Discord server messages to date: more than 2.05 billion messages from 4.74 million users across 3,167 servers.

desk verdict A genuinely useful, large Discord dataset that needs a few reproducibility fixes before I'd trust the headline numbers. read the letter →

arxiv 2502.00627 v1 pith:KUGICGC7 submitted 2025-02-02 cs.SI cs.DB

Yan Aquino , Pedro Bento , Arthur Buzelin , Lucas Dayrell , Samira Malaquias , Caio Santana , Victoria Estanislau , Pedro Dutenhefner
show 7 more authors
Guilherme H. G. Evangelista Luisa G. Porfírio Caio Souza Grossi Pedro B. Rigueira Virgilio Almeida Gisele L. Pappa Wagner Meira Jr
This is my paper · ORCID
classification cs.SIcs.DB
keywords Discordpublicserverdatasetcomputationalsocialsciencedecentralizedmoderationmediadatabotactivitymultilingualcommunitiesanonymization
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper presents a new public corpus of Discord communication and claims it is the largest of its kind: 2,052,206,308 messages from 4,735,057 users across 3,167 public servers, roughly 10% of servers listed in Discord's Discovery directory. The dataset spans from Discord's 2015 launch to December 2024 and was collected through Discord's public API, anonymized by pseudonyms and hashing, and released as per-server JSON files. A sympathetic reader would treat this as a resource claim: if the sampling and collection pipeline holds, researchers gain a uniquely broad, decade-long window into user-moderated communities, bot activity, and multilingual public conversation on a platform that other social networks' API restrictions have left understudied. Preliminary descriptions show English-dominant but multilingual server populations, heavy bot presence (17% of messages), and themes extending from gaming to art, music, and socializing.

What carries the argument

The carried object is the dataset itself, built from three components: Discord's Discovery directory as the sampling frame of 31,673 public servers, Discord's public API for retrieving channel metadata and message histories, and an anonymization layer that replaces usernames with consistent pseudonyms and applies 12-character SHA-256 hashes to user and message IDs while preserving record linkage. The machinery does the work of turning a platform-access affordance, namely Discord's retroactive chat history for public channels, into a reusable research artifact with per-server JSON files and a separate server-metadata file.

What would settle it

Compare the member counts, creation dates, preferred locales, and category keywords of the 3,167 sampled servers against the full 31,673-server Discovery snapshot from November 17, 2024; a statistically significant divergence on any of these axes would contradict the paper's assumption that the sample characterizes Discord's public-server population.

Watch

Extended reading notes

Core claim

The paper's central claim is that Discord Unveiled is the most extensive public dataset of Discord server messages assembled so far, containing over two billion messages and nearly five million distinct users from more than three thousand servers, and that a random 10% sample of servers listed in Discord's Discovery feature can support general study of public Discord communities. The paper supports the scale claim by describing a collection pipeline: enumerate all Discovery servers via URL queries, sample 10%, fetch every accessible text channel and message through Discord's API, then anonymize by pseudonymizing usernames and hashing user and message IDs before reorganizing into per-server JSON files. It also argues the corpus is broadly useful because its servers are multilingual and thematically diverse, and because Discord's user-driven moderation and bot ecosystem offer phenomena that centralized platforms do not.

Load-bearing premise

The dataset's representativeness rests on the assumption that a random 10% sample of Discord's Discovery servers (which only lists servers with at least 1,000 members and no explicit NSFW content) represents public Discord as a whole.

Editorial extensions

If this is right

  • With 2.05 billion messages across a decade, the dataset gives researchers a common substrate for studying how user-driven moderation evolves as communities grow.
  • The 17% bot-message share and the top-bot tables make it possible to quantify how moderation, entertainment, and gambling bots structure interaction on Discord.
  • The language distribution and keyword table support cross-linguistic and cross-thematic comparisons, for example between English- and Spanish-language public servers.
  • Because Discord's API makes historical public channels readable, the collection window in December 2024 to January 2025 still yields messages back to May 2015, allowing longitudinal studies that do not require years of prospective data collection.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: because the paper does not report a seed or random-number generator for its 10% selection, a reader who wants to generalize beyond these 3,167 servers should first test whether the sample matches the full Discovery list on member counts, creation dates, locales, and categories.
  • Editorial inference: the Discovery sampling frame excludes servers with explicit sexual content and servers with fewer than 1,000 members, so findings from this corpus should be scoped to Discovery-eligible public servers rather than all public Discord.
  • Editorial inference: a direct next test would be to run language identification on sampled message contents and compare it with the server-level preferred-locale field, since the paper only reports language at server metadata level.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 8 minor

Summary. The paper introduces a dataset of public Discord communication consisting of 2,052,206,308 messages from 4,735,057 users across 3,167 public servers, sampled as 10% of servers listed in Discord's Discovery feature. The authors describe a three-phase pipeline (collection via Discord's public API, anonymization via hashing and pseudonyms, and organization into server-level JSON files) and present descriptive statistics on temporal activity, bot usage, server languages, and server keywords. The central claim is that this is the most extensive Discord public server dataset to date, spanning 2015-2024.

Significance. If the dataset's completeness and representativeness are established, this would be a uniquely large and general-purpose corpus for studying decentralized moderation, online community governance, discourse, and bot ecosystems, complementing existing datasets from Reddit, Twitter/X, and Telegram. The paper provides a useful high-level description, detailed field documentation in the appendix, and a thoughtful anonymization scheme that preserves cross-message linkage. However, the scientific value of the contribution depends critically on the reproducibility and completeness of the collection process, which the current manuscript does not sufficiently demonstrate; the headline message count and the 2015-2024 temporal span are asserted rather than independently verifiable from the information given.

major comments (4)
  1. [Data Collection] The paper states that 'for each identified channel, we accessed all associated text messages' via the messages endpoint, but it does not describe the pagination logic (the endpoint returns at most 100 messages per call and requires a 'before' cursor), the authentication method (bot token vs. user token), whether the Message Content Intent was enabled, rate-limit/429 handling and retries, or how channels lacking read_history permission were treated. It also does not state whether threads and forum posts were included. Without this information and the collection code, the figure of 2,052,206,308 messages cannot be audited; systematic truncation would directly invalidate the 'most extensive' and 2015-2024 coverage claims. The authors should release the full collection script and provide internal validation checks, such as comparing the oldest message per channel with the server's creation date and re-fetching a subset of channels to show count stability.
  2. [Data Collection] The 'random 10% sampling' is not reproducible because no seed, random number generator, or algorithm is given. Furthermore, the sampling frame (Discord's Discovery tab) excludes servers with fewer than 1,000 members and servers that fail Discovery guidelines, so the sample is only representative of Discovery-listed servers, not of all public Discord servers. The manuscript's characterization of the sample as 'representative' should be explicitly qualified. Specify the exact sampling procedure and seed, and restrict all population-level claims to the Discovery frame.
  3. [Anonymization Process] Truncating SHA-256 to 12 characters produces a 48-bit hash space, for which the birthday bound among 4.7M users is on the order of 16 million, implying a nontrivial collision risk. The method for generating 'consistent pseudonyms' is also under-specified: if the pseudonyms are assigned deterministically from original usernames, the algorithm and seed must be reported; if they are not deterministic, cross-message linkage is broken. The regex-based replacement of user IDs embedded in message content is not validated for Discord's mention formats (e.g., <@!123>). Provide collision estimates, the precise pseudonymization algorithm, and a validation of the regex replacement.
  4. [Dataset Characterization] The abstract and introduction claim this is 'the most extensive Discord public server's data to date,' but the paper does not report the sizes of the prior Discord datasets it cites (Singh et al. 2024; Fillies, Peikert, and Paschke 2024). A quantitative comparison of message counts, server counts, user counts, and temporal spans is needed to substantiate the headline claim. Without such a comparison, 'most extensive' remains an assertion rather than a demonstrated property.
minor comments (8)
  1. [Data Availability] The DOI '10.5281/zenodo.146585059' and the URL in footnote 9 ('https://zenodo.org/records/14658505') do not match; the DOI should be verified and corrected, as this directly affects findability.
  2. [Figure 3] The caption for the bottom-right panel says the highlighted period is 'June 2022 to January 2024', but the text states the dataset spans through December 17, 2024 and collection ended January 2, 2025. Update the caption or the text to resolve the inconsistency.
  3. [Data Description] The phrase '2,052,206,308 unique messages' is misleading because message IDs are unique by construction; clarify whether any deduplication was performed during collection and, if so, how.
  4. [Table 1] The bot name 'Y AGPDB.xyz' appears to be a formatting error; the intended name is likely 'YAGPDB.xyz'.
  5. [Dataset Characterization] The language analysis uses the server-level `preferred_locale` field, which is a proxy for the language of actual messages; the paper should state this limitation more prominently, especially since the abstract claims 'substantial representations' of Spanish, French, and Portuguese.
  6. [Anonymization Process] The mimesis library is mentioned without a citation or version; provide a reference and version so the pseudonym generation is reproducible.
  7. [Paper Checklist] The checklist states that a Datasheet for the Dataset was created, but no datasheet or link is provided in the manuscript; include it as supplementary material or specify where it is hosted.
  8. [Throughout] There are several typographical and formatting issues, including 'Discord public server's data' in the abstract, 'individualJSON files' missing a space, 'V oice' in the Discord overview, and 'T urkish' in Figure 4.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: this is an observational dataset paper with no fitted parameters or derived predictions to reduce to inputs.

full rationale

The paper's contribution is a collected dataset plus descriptive summaries. There is no model, no fitted parameter, no uniqueness theorem, and no predictive claim that could be forced by construction. The central claims (2,052,206,308 messages, 4,735,057 users, 3,167 servers) are counts of collected observations, not outputs of a derivation. The only invocation of prior work by overlapping authors is the citation of Bento et al. 2024 to say that the data collection practices 'align with practices described in the current literature'; this is a methodological norm citation, not load-bearing evidence for the dataset's scale or composition. The paper's acknowledged limitations about Discovery-tab representativeness and server-side message deletion concern data completeness and external validity, not circularity: the dataset is what it is regardless of those caveats. Absent collection-code details and reproducibility artifacts are correctness/transparency risks, not circular reasoning. No step in the paper reduces an output to an input by definition, and no equational or definitional equivalence is present. Accordingly, the appropriate circularity score is 0.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The dataset paper introduces no new entities. Its claims rest on domain assumptions about Discord's API completeness, the representativeness of the Discovery-based sample, and the efficacy of the anonymization. The free parameters are methodological choices (sampling fraction, hash truncation length) rather than fitted values.

free parameters (2)
  • sampling_fraction = 0.1
    10% of Discovery servers selected via 'random sampling'; chosen for computational and storage constraints, not fitted to data.
  • hash_truncation_length = 12
    SHA-256 hashes truncated to 12 characters; chosen as a balance between anonymity and collision risk, not fitted to data.
assumptions (5)
  • domain assumption Discord's public API exposes all non-deleted historical messages in public channels.
    Stated in Data Characterization: 'Upon joining a server, users gain access to all non-deleted historical content within public channels, and the same is valid for data retrieval using their API.' Required for the 2015-2024 span.
  • domain assumption The 10% random sample of Discovery servers is representative of all public Discord servers.
    The paper generalizes trends (bots, languages, keywords) from the sample to Discord as a whole, but Discovery only lists servers with at least 1,000 members and compliant descriptions, so this is a strong assumption.
  • domain assumption SHA-256 truncation to 12 hex characters yields sufficiently low collision risk for reliable user and message linkage.
    The anonymization process truncates hashes to 12 characters (48 bits). With 4.7M users, collisions are possible, but the paper assumes it is sufficient for longitudinal analysis.
  • domain assumption Anonymization via pseudonyms and hashing prevents re-identification while preserving analytical utility.
    The paper claims this balance in Ethical Concerns, but also acknowledges 'no anonymization process can guarantee absolute anonymity'.
  • domain assumption The random sampling was implemented correctly with a random number generator, producing an unbiased sample.
    The paper only says 'random sampling' without specifying the method or seed, so the validity of the sample depends on an unverified implementation.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Discord Unveiled: A Comprehensive Dataset of Public Communication (2015-2024)." pith.science (2026). https://pith.science/paper/KUGICGC7

@misc{pith2026250200627,
  author       = {Pith},
  title        = {Pith review of: Discord Unveiled: A Comprehensive Dataset of Public Communication (2015-2024)},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KUGICGC7}},
  note         = {Machine review of arXiv:2502.00627}
}
read the original abstract

Discord has evolved from a gaming-focused communication tool into a versatile platform supporting diverse online communities. Despite its large user base and active public servers, academic research on Discord remains limited due to data accessibility challenges. This paper introduces 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. Preliminary analyses reveal significant trends in user engagement, bot utilization, and linguistic diversity, with English predominating alongside substantial representations of Spanish, French, and Portuguese. Additionally, prevalent community themes such as social, art, music, and memes highlight Discord's expansion beyond its gaming origins.

Figures

Figures reproduced from arXiv: 2502.00627 by the authors.

Figure 1
Figure 1. Discord interface, with the list of servers on the left, the distinct channels within the selected server, the central panel [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Diagram showcasing each step of the data collection process. [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Evolution of the daily number of messages sent over time. The top panel presents the complete time series, distin [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Bar plot of the number of servers by language, with [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Analyzing Political Discourse on Discord during the 2024 U.S. Presidential Election

    cs.SI 2025-02 conditional novelty 6.0 of 10

    A first look at political Discord servers shows Republican-aligned communities were more toxic and became more sexist after Kamala Harris's nomination, while Democratic servers focused on equality topics.

  2. Understanding Interpretation Difficulty in Harmful Online Communication: Insights from Cybercrime Communities

    cs.CL 2026-07 conditional novelty 5.0 of 10

    Interpreting harmful Discord messages requires integrating external knowledge and extended context, not just local message-level classification; LLMs leverage local context better than humans but still fail on coded l...

  3. Synopsis: Secure and private trend inference from encrypted semantic embeddings

    cs.CR 2025-05 conditional novelty 5.0 of 10

    A system combining local and central differential privacy with malicious-secure MPC lets journalists query semantic embeddings of donated E2EE messages without accessing the underlying texts.

Reference graph

Works this paper leans on

33 extracted references · 29 canonical work pages · cited by 3 Pith papers

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

  2. [2]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    T.; and Pignatiello, A

    Abi-Jaoude, E.; Naylor, K. T.; and Pignatiello, A. 2020. Smartphones, social media use and youth mental health. Cmaj, 192(6): E136--E141

    work page 2020

  4. [4]

    Ai, L.; Gupta, S.; Oak, S.; Hui, Z.; Liu, Z.; and Hirschberg, J. 2024. TweetIntent@Crisis: A Dataset Revealing Narratives of Both Sides in the Russia-Ukraine Crisis. Proceedings of the International AAAI Conference on Web and Social Media, 18(1): 1872--1887

    work page 2024

  5. [5]

    J.; and Rand, D

    Allen, J.; Watts, D. J.; and Rand, D. G. 2024. Quantifying the impact of misinformation and vaccine-skeptical content on Facebook. Science, 384(6699): eadk3451

    work page 2024

  6. [6]

    Aïmeur, E.; Amri, S.; and Brassard, G. 2023. Fake news, disinformation and misinformation in social media: a review. Social Network Analysis and Mining, 13(1): 30

    work page 2023

  7. [7]

    Bakshy, E.; Rosenn, I.; Marlow, C.; and Adamic, L. 2012. The Role of Social Networks in Information Diffusion. arXiv:1201.4145

    work page Pith review arXiv 2012

  8. [8]

    Balasubramanian, A.; Zou, V.; Narayana, H.; You, C.; Luceri, L.; and Ferrara, E. 2024. A Public Dataset Tracking Social Media Discourse about the 2024 U.S. Presidential Election on Twitter/X. arXiv:2411.00376

    arXiv 2024

Show all 33 references
  1. [9]

    M.; Tekumalla, R.; Wang, G.; Yu, J.; Liu, T.; Ding, Y.; Artemova, K.; Tutubalina, E.; and Chowell, G

    Banda, J. M.; Tekumalla, R.; Wang, G.; Yu, J.; Liu, T.; Ding, Y.; Artemova, K.; Tutubalina, E.; and Chowell, G. 2023. A large-scale COVID-19 Twitter chatter dataset for open scientific research - an international collaboration

    2023

  2. [10]

    Baumgartner, J.; Zannettou, S.; Keegan, B.; Squire, M.; and Blackburn, J. 2020. The Pushshift Reddit Dataset. arXiv:2001.08435

    2020 arXiv

  3. [11]

    Bento, P.; Buzelin, A.; Aquino, Y.; Carvalho, I.; Dutenhefner, P.; Dayrell, L.; Santana, C.; Estanislau, V.; Pappa, G.; Miranda, D.; Almeida, V.; and Jr, W. M. 2024. Impacto da Pandemia na Discussão sobre Saúde Mental: O Caso do Discord no Brasil. In Proceedings of the 30th Br...

    2024

  4. [12]

    Dooms, S.; De Pessemier, T.; and Martens, L. 2013. Movietweetings: a movie rating dataset collected from twitter. In Workshop on Crowdsourcing and human computation for recommender systems, CrowdRec at RecSys, volume 2013, 43

    2013

  5. [13]

    Efstratiou, A.; Blackburn, J.; Caulfield, T.; Stringhini, G.; Zannettou, S.; and De Cristofaro, E. 2023. Non-polar Opposites: Analyzing the Relationship between Echo Chambers and Hostile Intergroup Interactions on Reddit. Proceedings of the International AAAI Conference on Web...

    2023

  6. [14]

    Fillies, J.; Peikert, S.; and Paschke, A. 2023. Hateful Messages: A Conversational Data Set of Hate Speech produced by Adolescents on Discord

    2023

  7. [15]

    Fillies, J.; Peikert, S.; and Paschke, A. 2024. Hateful Messages: A Conversational Data Set of Hate Speech Produced by Adolescents on Discord. In Haber, P.; Lampoltshammer, T. J.; and Mayr, M., eds., Data Science---Analytics and Applications, 37--44. Cham: Springer Nature Swit...

    2024

  8. [16]

    Gillespie, T. 2018. Custodians of the Internet: Platforms, Content Moderation, and the Hidden Decisions That Shape Social Media. ISBN 9780300235029

    2018

  9. [17]

    Hoevers , S. 2022. Discourse on Discord : An analysis of relations between users

    2022

  10. [18]

    K.; and Salter, A

    Johnson, E. K.; and Salter, A. 2022. Embracing discord? The rhetorical consequences of gaming platforms as classrooms. Computers and Composition, 65: 102729

    2022

  11. [19]

    Kiene, C.; and Hill, B. M. 2020. Who Uses Bots? A Statistical Analysis of Bot Usage in Moderation Teams. In Extended Abstracts of the 2020 CHI Conference on Human Factors in Computing Systems, CHI EA '20, 1–8. New York, NY, USA: Association for Computing Machinery. ISBN 9781450368193

    2020

  12. [20]

    Kraut, R.; and Resnick, P. 2012. Building Successful Online Communities: Evidence-Based Social Design. ISBN 9780262298315

    2012

  13. [21]

    La Morgia, M.; Mei, A.; and Mongardini, A. M. 2023. TGDataset

    2023

  14. [22]

    Leetaru, K.; and Schrodt, P. A. 2013. GDELT: Global data on events, location, and tone. ISA Annual Convention

    2013

  15. [23]

    Lewis, K.; Kaufman, J.; Gonzalez, M.; Wimmer, A.; and Christakis, N. 2008. Tastes, ties, and time: A new social network dataset using Facebook. com. Social networks, 30(4): 330--342

    2008

  16. [24]

    Mekacher, A.; Falkenberg, M.; and Baronchelli, A. 2024. The Koo Dataset: An Indian Microblogging Platform with Global Ambitions. Proceedings of the International AAAI Conference on Web and Social Media, 18(1): 1991--2002

    2024

  17. [25]

    Pater, J.; Kim, S.; Mynatt, E.; and Fiesler, C. 2016. Characterizations of Online Harassment: Comparing Policies Across Social Media Platforms. In Proceedings of the 19th International Conference on Supporting Group Work (GROUP '16), 369--374. New York, NY, USA: Association fo...

    2016

  18. [26]

    Pierri, F.; Luceri, L.; Jindal, N.; and Ferrara, E. 2023. Propaganda and Misinformation on Facebook and Twitter during the Russian Invasion of Ukraine. In Proceedings of the 15th ACM Web Science Conference 2023, WebSci '23, 65–74. New York, NY, USA: Association for Computing M...

    2023

  19. [27]

    Seering, J. 2020. Reconsidering Self-Moderation: the Role of Research in Supporting Community-Based Models for Online Content Moderation. Proc. ACM Hum.-Comput. Interact., 4(CSCW2)

    2020

  20. [28]

    Seering, J.; Wang, T.; Yoon, J.; and Kaufman, G. 2019. Moderator engagement and community development in the age of algorithms. New Media & Society, 21(7): 1417--1443

    2019

  21. [29]

    Seufert, A.; Poignée, F.; Hoßfeld, T.; and Seufert, M. 2023. WhatsApp Data Set

    2023

  22. [30]

    Shang, L.; Chen, B.; Vora, A.; Zhang, Y.; Cai, X.; and Wang, D. 2024. SocialDrought: A Social and News Media Driven Dataset and Analytical Platform towards Understanding Societal Impact of Drought. Proceedings of the International AAAI Conference on Web and Social Media, 18(1)...

    2024

  23. [31]

    K.; Ghafouri, V.; Such, J.; and Suarez-Tangil, G

    Singh, A. K.; Ghafouri, V.; Such, J.; and Suarez-Tangil, G. 2024. Differences in the Toxic Language of Cross-Platform Communities. Proceedings of the International AAAI Conference on Web and Social Media, 18(1): 1463--1476

    2024

  24. [32]

    L.; Luceri, L.; Pierri, F.; and Ferrara, E

    Wang, E. L.; Luceri, L.; Pierri, F.; and Ferrara, E. 2023. Identifying and Characterizing Behavioral Classes of Radicalization within the QAnon Conspiracy on Twitter. arXiv:2209.09339

    2023 arXiv

  25. [33]

    A.; and Land, M

    Wilson, R. A.; and Land, M. K. 2020. Hate speech on social media: Content moderation in context. Conn. L. Rev., 52: 1029

    2020

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.