YouTube - Social Network Analysis on Amber Heard's Case Example from Data Analysts, Researchers, and Scientists on the Disinformation/Influence Operation.
- E.g., Videos, Accounts, Multiple Accounts with Same Comments, Timings, Like/Dislike Ratios, Coordinated Inauthentic Activity, Cluster Networks
YouTube - Social Network Analysis on Amber Heard's Case Example from Data Analysts, Researchers, and Scientists on the Disinformation/Influence Operation.
The Disinformation Operation is highly voluminous, blatant, and threatening on YouTube.
- E.g., Videos, Accounts, Multiple Accounts with Same Comments, Timings, Like/Dislike Ratios, Coordinated Inauthentic Activity, Cluster Networks
Data Collections:
YouTube: (1) 2,176,926 Comments (1,693,341 Comments with 483,585 Replies), 6,893 Videos (with dislike ratios), 1,152,868 Accounts, Cross-Platform Links
At minimum Tens of Thousands of Repeated Same/Exact Comments...
Higher amounts of comments against Amber Heard correlate with increased threats seen in Heat Maps. Correlate to dangers in-person.
Threat Analysis is shown above of comments with:
Threat - 149,810 - 6.88%
Hate-Speech - 351,399 - 16.1%
Others - 1,675,334 - 77%
Threat Keywords:
Kill & Death 52,232 - 33.2%
Crime & Jail 75,830 - 48.2%
Hell & Burn 29,214 - 18.6%
High Volume of Repeated Exact Comments - Unlike as high volume as other platforms, on YouTube the exact phrases blatantly repeat with bot-like actions (exception: Instagram 136 times of "i am inevitable" with 136 accounts in May 2019). YouTube removed dislikes on their entire platform around a year after the start of the "Amber Heard's 2020 Takeaway: Adapt and Survive" with same exact comments and 479K dislikes. 4 other videos, e.g., Gully trailer with 92K dislikes and settlement video receive higher dislikes.
The operations focus on many accounts, wide spread, less so the same accounts, yet, even on YouTube, there are praising texts blatantly using the same usernames pointing to the source and are of the same accounts. Each platform serves different purposes.
Some of Most Repeated Comments
df_comments.text.value_counts().head(30)
they won't believe you because you are a man she is not a victim 43370
that's crazy 5968
woah what 2797
ًں’“
thanks for your utmost love and support which has brought me this far it has been a hard time for me but your support has got me standing even in difficult times i can't appreciate enough the support of fans who have been really active with issues relating to meprivate message on telegram @johnnydeppllofficial ًں’“ much love â¤ï¸ 2786
they won't believe you because you are a man amber heard 2253
lol 1705
they won't believe you because you are a man amber heard she is not a victim 1585
wow 1501
1399
they won't believe you because you are a man 1352
they won't believe you because you are a man she is not the victim 1343
ں’•ًں’“thanks for your utmost love which has brought me this far it has been a hard time for me but your support got me going where are you watching from?ًں’• 1220
hi 1134
she is not a victim 1110
they wont believe you because you are a man she is not a victim 1020
they won't believe you because your a man she is not a victim 1012
i love this song 951
yes 944
they won't believe you because you are a man she is not a victim! 929
nice 922
wtf 892
same 823
me 797
whoa what 795
that's crazy! 763
exposed 753
good 730
Name: text, dtype: int64
Negative Comments Peak Dates:
Dec 27, 2020 - 14113
Dec 23, 2020 - 11688
Dec 26, 2020 - 10880
Dec 24, 2020 - 10446
Dec 28, 2020 - 9164
More data collections being conducted of 18K videos on AH (search terms given).
Deeper Analysis and Relational Dataframe is being worked on. Note: Anyone can obtain YouTube APIs.