Raiders of the Lost Kek: 4chan /pol/ Labeled Posts (2016-2019)

A 4chan /pol/ (Politically Incorrect board) dataset of 134,529,233 posts across 3,397,911 threads (4,894,179 with 7-score Perspective API annotations), collected by Papasavva et al. via 4chan API + Perspective API (2020). Each record contains post text (com), thread ID, post number, country, timestamp, image metadata, 7 Perspective API scores (TOXICITY, SEVERE_TOXICITY, INFLAMMATORY, PROFANITY, INSULT, OBSCENE, SPAM), named entities. 134,529,233 4chan /pol/ posts with per-post Perspective API toxicity scores across 7 dimensions — one of the most comprehensively annotated fringe platform datasets covering 3.5 years of /pol/ discourse. Suitable for toxicity measurement, hate speech analysis, fringe platform research, political discourse on imageboards, cross-platform toxicity comparison, longitudinal hate speech trends.

Very large: 134.5M records. Working with all of it takes more than a laptop (tens of GB of memory or a disk-based graph database); a sample, or reading the files in a streaming way, is the usual way to start.

#4chan pol#Fringe platform#Imageboard#Named entity#Perspective api

Download data BibTeX JSONSimilar datasets

Analytics

Content metrics

4,894,179 records scored · min–maxmean ± 0.675 σ (the interquartile range if the values are roughly normal)mean

toxicityμ 0.42 ± 0.309
min 0max 0.997
low 33%medium 34%high 33%
severe toxicityμ 0.315 ± 0.277
min 0.001max 0.985
low 33%medium 34%high 33%
inflammatoryμ 0.408 ± 0.246
min 0max 0.994
low 33%medium 34%high 33%
profanityμ 0.362 ± 0.328
min 0.001max 1
low 33%medium 34%high 33%
insultμ 0.391 ± 0.289
min 0.001max 1
low 33%medium 34%high 33%
obscene37.0% true
true 37.0%false 63.0%
spam19.2% true
true 19.2%false 80.8%
8.0
Flesch–Kincaid grade
US school grade level
23.3
Automated readability index
Frequent terms
classhrefquotelinkspanquotelikepeopledonfuckingshitfucktrumpthinkgoodknowrightwantwhitecountrytimepolgoingworldactuallyneedwaygotdoesnthreadgodwomenisnlooknewyearsjewsmanthingbetterliterallysaidworkdidnstopliveprettymoneybadkilllife
Topic clusters

Topic 0

trump
hillary
poll
clinton
vote
shill
win
election

Topic 1

jews
israel
jew
jewish
palestine
israeli
holocaust
palestinian

Topic 2

gun
kill
shoot
weed
fire
child
police
cop

Topic 3

degree
tier
debt
pay
job
education
student
business

Topic 4

god
religion
logic
hinduism
morality
true
moral
believe

Topic 5

muslim
isis
islam
terrorist
muslims
kill
islamic
religion

Topic 6

nice
pic
good
lad
guy
relate
look
great

Topic 7

capitalism
socialism
libertarian
economic
market
breach
socialist
free

Topic 8

dog
bark
cat
animal
pet
dogs
owner
abuse

Topic 9

hitler
germany
poland
german
germans
europe
aryan
european

Topic 10

faggot
dick
suck
gay
cock
porn
fuck
fucking

Topic 11

sweden
swedish
swede
finland
dutch
woman
norway
iceland
Show all 36 topics

Topic 12

span
store
named
image
write
status
thread
replace

Topic 13

nigger
white
hate
kid
fuck
black
dick
kill

Topic 14

boris
gove
corbyn
tory
party
labour
leadership
run

Topic 15

fill
spill
florida
north
japan
expand
texas
dakota

Topic 16

girl
look
love
cute
face
ugly
hot
like

Topic 17

know
noice
digit
question
answer
right
accurate
number

Topic 18

russia
slavs
russian
slavic
slav
soviet
putin
space

Topic 19

fuck
cunt
peter
fetish
fucking
pete
christopher
fucker

Topic 20

think
happen
coincidence
thinking
wonder
thing
olympic
wow

Topic 21

meme
retweet
tumblr
twitter
newfag
totally
neat
tweet

Topic 22

read
book
forget
account
lie
deposit
dislike
library

Topic 23

kek
lord
praise
approve
moloch
powerful
keked
lament

Topic 24

shit
shame
disgusting
rude
brutal
christ
mad
wtf

Topic 25

brexit
leave
referendum
britain
vote
gove
british
bro

Topic 26

usa
country
america
canada
ontario
american
england
kentucky

Topic 27

video
watch
stream
hear
sound
immediate
nervous
viewer

Topic 28

month
time
day
map
week
end
post
join

Topic 29

base
senpai
ahmed
desu
man
calm
shiva
japa

Topic 30

toilet
paper
flush
poo
seasoned
iron
cast
kebab

Topic 31

leaf
anon
leadsom
burn
fucking
good
tree
character

Topic 32

het
dat
que
aan
van
dit
naar
hoe

Topic 33

exercise
calorie
weight
eat
fat
lose
diet
muscle

Topic 34

funny
laugh
boom
joke
hilarious
chuckle
comedian
nig

Topic 35

hate
woman
feminism
feminist
beta
fucking
abortion
wtf
Top hashtags
#p79291774101
#p7929962176
#p7928902668
#p7924380364
#p7927352362
#p7924956061
#p7927772758
#p7929323157
#p7927819155
#p7929678852
#p7927386450
#p7929071048
#p7928222446
#p7928897045
#p7927515442
#p7929128742
#p7929369738
#p7929604838
#p7928839737
#p7929311337

Computed over the full corpus.

Record

Catalog key
4chan-pol-labeled
Authors
Papasavva, Antonis; Zannettou, Savvas; De Cristofaro, Emiliano; Stringhini, Gianluca; Blackburn, Jeremy
Year
2020
Source paper
Raiders of the Lost Kek: 3.5 Years of Augmented 4chan Posts from the Politically Incorrect Board
DOI / arXiv
10.1609/icwsm.v14i1.7354
License
CC BY 4.0 (CC-BY-4.0)
Records
134,529,233
Languages
en
Collected
2016-06 to 2019-11 (2010s)
Tasks
Toxicity detection (regression; toxicity, severe_toxicity, insult, obscene) inferred
Region
Global (inferred: the platform is used worldwide)

Provenance

Originally published at
https://zenodo.org/records/3606810 link checked
Hosted copy
Mirrored on echo.directory on 2026-10-02, 22.4 GB across 1 files. Checksums in MANIFEST.json.

How to cite

BibTeX

@misc{echo2026,
  title  = {ECHO: Exploratory Compendium of Humanity Online},
  author = {Switzer, Braden and Zafarani, Reza},
  year   = {2026},
  url    = {https://echo.directory},
  note   = {Syracuse University}
}

@article{Papasavva_2020,
  title     = {Raiders of the Lost Kek: 3.5 Years of Augmented 4chan Posts from the Politically Incorrect Board},
  volume    = {14},
  issn      = {2162-3449},
  url       = {http://dx.doi.org/10.1609/icwsm.v14i1.7354},
  doi       = {10.1609/icwsm.v14i1.7354},
  journal   = {Proceedings of the International AAAI Conference on Web and Social Media},
  publisher = {Association for the Advancement of Artificial Intelligence (AAAI)},
  author    = {Papasavva, Antonis and Zannettou, Savvas and De Cristofaro, Emiliano and Stringhini, Gianluca and Blackburn, Jeremy},
  year      = {2020},
  month     = May,
  pages     = {885–894}
}

APA

Switzer, B., & Zafarani, R. (2026). ECHO: Exploratory Compendium of Humanity Online [Data repository]. Syracuse University. https://echo.directory

Papasavva, A., Zannettou, S., De Cristofaro, E., Stringhini, G., & Blackburn, J. (2020). Raiders of the Lost Kek: 4chan /pol/ Labeled Posts (2016-2019) [Data set]. https://doi.org/10.1609/icwsm.v14i1.7354

MLA

Switzer, Braden, and Reza Zafarani. ECHO: Exploratory Compendium of Humanity Online. Syracuse University, 2026, echo.directory.

Papasavva, Antonis, et al. Raiders of the Lost Kek: 4chan /pol/ Labeled Posts (2016-2019). 2020, doi.org/10.1609/icwsm.v14i1.7354.

BibTeX · RIS · CSL-JSON

Similar datasets 8 all with scores →