#MemeWar network visualizations

Erin Gallagher
6 min readSep 30, 2018

Data visualizations of 20,000 tweets from the #MemeWar against CNN in July 2017.

User-to-hashtag network for 20,072 #MemeWar tweets from 07/05/2017 to 07/12/2017

This is blog # 7 in a series of blogs about 4chan trolling tactics during the 2016 election. View the entire series here.

Network graphs in this blog represent a sample of 20,072 #MemeWar tweets I downloaded via TweetArchivist from July 5 to July 12, 2017.

Operation Autism Storm, also known as the Great CNN Meme War, is a 4chan operation urging participants to spread anti-CNN image macros, videos and animated GIFs in retaliation for the news network threatening to dox a Redditor for posting a GIF of Donald Trump bodyslamming the CNN logo in early July 2017. — Know Your Meme

I went back through the dataset I collected and found a large number of 3rd party apps were used to tweet in #MemeWar. Besides the usual common sources (iPhone, Android, Web Client, etc), here is the list of apps I found in this dataset:

Bot Libre!
Visual Decolonization Tweet Bot
DMI Summer School 2017
Sakpol Magic
Twidere for Android #7
Dank Retweets
Pruebas Pythin
TwitPane Android
Tweetium for Windows
Flamingo for Android
Ze Memes
new app kobeck
Top Email Retweets
Harambe Bot App
EV App
Crowdfire — Go Big
UberSocial for BlackBerry
Blaq for BlackBerry® 10
WWE News and Gossip Feed App
Conservative Riposte
TwitterDub @Devnullius
Talon (Plus)
Fenix for Android
Tweetbot for iΟS
only math

Some of these are legit Twitter services like, TweetDeck, IFTTT and RoundTeam, others are custom built apps. The sheer amount of apps in one hashtag is remarkable to me. It’s normal to see one or two other apps besides the usual sources but this many apps is odd.

User-to-hashtag network for 20,072 #MemeWar tweets from 07/05/2017 to 07/12/2017

Fake accounts

Accounts that tweeted this hashtags were obviously fake and laden with alt-right and neo-nazi references. I randomly chose one account of several accounts that was tweeting using Buffer.

kendrickek77 was created on July 8, 2017, tweeted a total of 16 tweets. “Kendrick E. Kornes” (aka “KEK”) is a German fake account that participated in the meme war against CNN. They tweeted 15 #MemeWar tweets in English language but the first tweet they retweeted and liked is in German language and they followed 79 German accounts.

Here’s a list of other accounts that were using Buffer in addition to kendrickek77

kendrickek77’s few tweets have anti-semitic tropes and neo-nazi references.

One of hundreds of GIFs of Trump beating CNN; this was taken from a scene in the movie American History X where Edward Norton, playing a white power skinhead “curb stomps” an African American man (ie: makes him bite the curb then stomps on the back of his head, violently murdering him).

Filtering the network by degree range 25 reveals some highly active accounts as well as a variety of other hashtags that were used with #MemeWar.

User-to-hashtag network for 20,072 #MemeWar tweets from 07/05/2017 to 07/12/2017 filtered by degree range 25

There were also many fake, throw-away accounts that were created between July 5 and July 8— used only for the #MemeWar trends and have not tweeted since such as CNNISISIS4, trumpbeatscnn, MemeWarSoldier, TrumpAndCNN and FuckCNNtohell (screenshot below). A few other obvious fake accounts appear to have deactivated such as anticnnmemes, CNNisMad, CNNMeme_Warrior and DownWithCNN.

Fake account created on July 6, 2017 for #MemeWar and has not tweeted since

General stats from Tweet Archivist on #MemeWar

Influencer Index — accounts with the highest number of followers that tweeted or retweeted tweets containing hashtag #MemeWar. The first account that stands out to me is AppSame, the conservative political marketing firm that was also in the Inlfuencer Index for #QAnon, #WalkAway, #TheStorm as well as several high volume accounts I recently documented in #QAnon hashtags. AppSame had over 1 million followers in 2017 so it has lost a significant number of followers since the Great CNN Meme War. I also recognize several other high volume accounts that I found in #QAnon trends like LeahR77, SandraTXAS and GaetaSusan.

Hashtags that were linked to #MemeWar

#MemeWar — Volume Over time

The tweets in my sample dataset made 58,742,745 impressions per Tweet Archivist who defines the metric as “the total number of times that the tweets of an archive have been delivered to Twitter streams. Of course, not everyone who receives a tweet will read it. As such, impressions are the largest possible audience for the given archive. Paid advertising works similarly; even though an ad was displayed on a website, there is no guarantee that a person actually saw it.”

User-to-user network

Here’s the user to user network for the same dataset of 20,072 #MemeWar tweets. It contains 12,900 nodes, 18,909 edges and 617 communities.

User-to-user network for 20,072 #MemeWar tweets from 07/05/2017 to 07/12/2017

The largest node, jojoh888 only tweeted 13 tweets with the hashtag #MemeWar over a period of 5 days but their tweets went viral which is why their node appears massive. jojoh888 only had around 40K followers at the time so they are not included in the Influencer Index either but they were clearly one of the main accounts that received the most diffusion in this hashtag.

CNN and Donald Trump’s accounts appear as central nodes because many people were mentioning both of those handles in tweets. When a handle is mentioned excessively, it makes the node appear as if it’s being stretched in all directions (literally being dragged in the case of CNN’s account), as opposed to jojoh888’s account which looks more like a burst.

User-to-user network for 20,072 #MemeWar tweets from 07/05/2017 to 07/12/2017

The orange cluster was the Infowars crew: The main Infowars account, RealAlexJones and PrisonPlanet as well as Jack Posobiec, Stefan Molyneux, Mark Dice and Baked Alaska.

User-to-user network for 20,072 #MemeWar tweets from 07/05/2017 to 07/12/2017

Filtering the user-to-user network by degree range 25 shows some of the more active accounts in this hashtag.

User-to-user network for 20,072 #MemeWar tweets from 07/05/2017 to 07/12/2017 filtered by degree range 25

Given the long list of apps I found in this dataset, I can say there was definitely some level of automation in the 2017 hashtag #MemeWar. My guess is there was also a lot of sock puppetry too. Referring back to “Connecting Social Media and Pushing Content 101 & 102” it’s a reasonable assumption that 4chan trolls who used Android emulators to run multiple accounts during the election may have continued using the same tactics after the election.

Next Up:



Erin Gallagher

Social media researcher, multimedia artist, former research assistant with the Technology and Social Change Project