Two days prior to my account being suspended, I tweeted the following (note several Twitter employees were copied):
The @LisaMar91564392 account — which boasts a background photo that prominently displays “Trudeau Must Go”— should have been suspended long ago for engaging in coordinated, inauthentic activity:
Dissecting “Lisa the Nationalist for Canada”
Below is a network graph representation of 41,085 Twitter accounts (where @LisaMar91564392 is the most central account):
In addition to collecting connections data (Following/Followers relationships), Social Forensics collected profile data (tweet count, like count, account creation date, bio, url, display name, location, etc.) for the 41,085 accounts.
The graph contains three colors (light green, blue, dark green). Each color represents an algorithmically determined (modularity class) community. An account’s community is based on (Following/Followers) interconnectivity. In other words, an account that appears as a blue node will tend to be Following and have more Followers from the blue community than from the light/dark green communities.
Next, we’ll visualize the data after isolating ten (keyword-driven) segments. The segments are as follows: 1) Brexit, 2) Canada, 3) KAG, 4) MAGA, 5) Patriot, 6) QAnon, 7) Trudeau, 8) Trump, 9) Wexit and 10) WWG1WGA
Accounts are included in segments should at least one of three conditions apply: 1) keyword appears in account’s username, 2) keyword appears in account’s display name or 3) keyword appears in account’s bio
1) Brexit (674 accounts)
2) Canada (768 accounts)
3) KAG (10,724 accounts)
4) MAGA (18,778 accounts)
5) Patriot (6,885 accounts)
6) QAnon (1,613 accounts)
7) Trudeau (653 accounts)
8) Trump (14,882 accounts)
9) Wexit (139 accounts)
10) WWG1WGA (4,281 accounts)
Using this basic keyword-driven segmentation/visualization, we can quickly get a feel for the types of accounts that comprise each community. Generally speaking, Canadian-focused accounts are on the right (dark green community) while MAGA-focused accounts are on the left (blue and light green communities). Additionally, there’s a much smaller UK-focused sub-community within the blue community (bottom center; see Brexit segment).
Here’s a summary of the ten keyword-driven segments (most notably, 46% of the 41,085 accounts contain the word “MAGA” in their username, display name or bio!):
Bridging MCGA (Make Canada Great Again) & MAGA
As we saw earlier, “Lisa the Nationalist for Canada” functions as a bridge between anti-Trudeau (nationalist types) and MAGA focused communities:
Of the 975 accounts that “Lisa the Nationalist for Canada” is Following, 815 (84%) are Followers of @LisaMar91564392 (across the account’s overall Following, 3.7K accounts, that becomes 81%). Translated: the account largely built its audience in a follow-for-follow type fashion
Several other accounts similarly function as bridges between Canadian (right) and United States (left) focused communities. Here are a few examples (note how each account has a larger Following than Followers count):
As is the case with @LisaMar91564392, these accounts (@He48300141, @lwrattail, @PaulSR36855907 and @PersonalJihad) similarly exist for the sole purpose of distorting the public debate. In fact, the same can be said for the vast majority of accounts that comprise the 41,085 account dataset.
Hyperactive Accounts (100+ Tweets/Day)
Highlighted below are accounts (2,188) that have, on average, tweeted 100+ times per day since being created (including @LisaMar91564392):
In aggregate, there have been 310 million tweets from the 2,188 accounts.
Within the context of the 41,085 account dataset, “Lisa the Nationalist for Canada” is Following 59 hyperactive tweeting accounts (top), while 64 hyperactive tweeting accounts are Followers of @LisaMar91564392 (bottom):
Alternatively visualized (and where labels have been added; Following=top and Followers=bottom):
The 59 hyperactive tweeting accounts (full usernames) that “Lisa the Nationalist for Canada” is Following can be found here and below:
Below I have reapplied visual clustering after filtering the dataset to include all (2,188) hyperactive tweeting (100+ tweets/day) accounts:
The 2,188 hyperactive tweeting accounts appearing in the 41,085 account dataset can be found here.
Flag Bearing Nationalist Types
I didn’t seek out the patriots; rather, they found me.
Thanks to Twitter suspending my account (@geoffgolberg) — alleging I was abusive towards “Lisa the Nationalist for Canada” (@LisaMar91564392) — Social Forensics has discovered what appears to be a massive, connected influence operations campaign that spans at least three countries (Canada, United Kingdom and United States).
To be clear, when stating the operations span at least three countries, this references the focus of the various astroturfing efforts rather than the physical locations of those behind said accounts.
Twitter claims the recent Canadian federal election was absent “any instances of large-scale attempts at disinformation or manipulation.”
Our analysis, on the other hand, reveals thousands of Canadian-focused accounts engaging in inauthentic activity that continue to weaponize Twitter’s platform via distorting the public debate.
Several of those accounts are cycled through below (after filtering to include only the dark green community and reapplying visual clustering):
It is Social Forensics’ belief that Big Tech should be audited (re: their authentic user base) in the same fashion that Big 4 accounting firms audit the financials of publicly traded companies:
“It is difficult to get a man to understand something, when his salary depends on his not understanding it.” -Upton Sinclair
Geoff Golberg is an NYC-based researcher (and entrepreneur) who is fascinated by graph visualization/network analysis — more specifically, when applied to social networks and blockchain activity. His experience spans structured finance, ad tech, and digital marketing/customer acquisition, both at startups and public companies.
Geoff is the Founder/CEO/Janitor of Social Forensics, where he spends (far too much of) his time developing techniques and building tools to identify social media manipulation (of various flavors!).
Read about Geoff’s war with Twitter here!