Skip to content

Instantly share code, notes, and snippets.

@salcoast
Forked from travisbrown/deznat-centrality.md
Created September 23, 2021 06:06
Show Gist options
  • Save salcoast/a78b5dbdcf3f795f5fba29f8f8c529de to your computer and use it in GitHub Desktop.
Save salcoast/a78b5dbdcf3f795f5fba29f8f8c529de to your computer and use it in GitHub Desktop.
Twitter accounts most prominently associated with #DezNat (Deseret Nationalism)

I've seen some reactions to this list (some people are mad that they're on it, some people are mad that they aren't on it, some people are mad that other people are on it, some people are mad that they aren't on it any more, some people are going to pay me a visit), and most of those reactions are based on misinterpretations, so I'm shortening the list and adding an explanation.

This list isn't a ranking of who is most ideologically committed to this hate group. I don't know who any of these people are. I'm "just making shit up" in the sense that I'm running a couple of fairly simple algorithms on some data I've collected. The goal I have for those algorithms is to identify "communities" (in a very abstract sense) on Twitter, and also to identify the accounts that are most "central" (also in an abstract sense) to those communities. Hypothetically someone could be "central" in this sense to the conversations that members of an extremist group are having on Twitter without being a member of the group themselves. The algorithm doesn't care.

In the case of these Deseret nationalists (who I'd never heard of before this month), I seeded the data collection step with a list of accounts from BlockDNat, a group that describes itself as " blocking alt-right & troll accounts in Mormon-adjacent Twitter". The data collection system (described in more detail here) then found other accounts whose interactions were primarily with accounts on the BlockDNat list. After a few iterations of that, the system produced a list of a couple thousand accounts, some of which probably don't actually have much to do with any kind of religious fundamentalism.

The next step is what produced this list. The system identifies three kinds of connections between accounts:

  • Retweets
  • Replies
  • Shared presence in a reply thread

The first two are weighted much more highly than the third. Note that the system doesn't use information about who follows who, since it's intended to provide a view of the entire history of a community on Twitter, and we don't have historical information about the follower graph.

These three kinds of connections between accounts give us a network graph, and we can then analyze that network with fairly off-the-shelf implementations of well-known algorithms, which is where this list came from.

None of this analysis is very sophisticated. I'm not a data scientist. My work on the more general project here (tracking far-right speech on Twitter) started because of a practical need but is mostly a hobby. Some other people have found it useful, so I've been publishing a few examples of its application (like this list).

If you're really proud of your involvement in this hate group and wish you were ranked more highly, all I can say is: produce better content. Focus on engagement. Maybe try not to be such a reply guy.

(One final thing to note is that this list is based entirely on publicly available data, all of which was sourced from the Wayback Machine, not the Twitter API or any Twitter client.)

Score Screen names Twitter ID
1 0.02703297 JPBellum 4745389657
2 0.01952807 JAEbberts 1081719290133442566
3 0.01833583 JReubenCIark 4852741342
4 0.01379722 Mormonger, colenoorda 1408886100
5 0.01269674 Neil10790465 886046676
6 0.01106221 StJohn724, MormonLibsLMAO 1029587993512824832
7 0.01072696 chhardman 26649679
8 0.01030932 Matthew_7_14, Matt____FL 1442170544
9 0.01021471 scottleish 397792123
10 0.01018112 65D_68W_4H0, girdedloins8 1237797424602570752
11 0.00925232 Bossynotbossy, FriedScones, WiseBossy, LiveWiseBossy, LiveWiseLifeNow 954551442244714496
12 0.00885642 NotYourBishop 1281051965636481025
13 0.00876541 tannerguzy, MasculineStyle 346063582
14 0.00865847 _154831 522598451
15 0.00851528 jcbonthedl, KMDontheUH 1025615726294384640
16 0.00802124 idahohurricane, Trinabased, TrinaFaye20, Katrinafaye 34443952
17 0.00795442 SpacePiolet, DeviousWing, NSteele1991 70709964
18 0.00770169 iamhomosexuaI, UtahValleyDaddy, NiasDiad, ElMioCid3 1245591829690478593
19 0.00762405 ZionBuild____, ZionBuild___, ZionBuild__, Build_Zion_, Build_Zion, Jeremy_Ephraim1 1113795464334155776
20 0.00722234 HannaSeariac 909106570906054657
21 0.00710867 larknap 1070553945896976385
22 0.00693654 TheTriarii, ParagonZerg 3002684195
23 0.00672140 darthcaro, disneecaro 1257248779276742656
24 0.00651614 Tower_Overwatch, stulta_diximus 260834188
25 0.00618562 BackFromThat 901126590117011456
26 0.00594283 hankrsmith 2404891854
27 0.00589562 zarahemla21, _KingBenjamin 1312418083433316352
28 0.00585759 ABouncingSoul, RooneyB21 1212420582836301824
29 0.00578265 RawHoneyBrah, BrettCainBooks 1172086334229372928
30 0.00558379 huddleb2, spinachtyrant99, Amish_tyrant99, spinachthot99, bhudd99, pitied_fool, barnett_stan99, beljeet_97, Muz_Mo99, MuzMo99 380620515
31 0.00546064 _154831 1326037483923013633
32 0.00512090 LeilandTanner 313774045
33 0.00511371 Jessla_G 1125953177742917632
34 0.00497316 Luv2GoFly 977914992296013826
35 0.00494988 bobdaduck 277536867
36 0.00486645 LatterdayDoofus 1020346707802902528
37 0.00483996 FiverMacGyver, BuddyMacGyver 1278585428
38 0.00480255 JesseLucasSaga 451292765
39 0.00466042 og_BYU, BYYYYYYYYYYU 845865030163742720
40 0.00459643 ConflictJustice, M__Game 47244501
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment