Friday, October 21, 2016

ISIS Twitter Analysis: The “Awlaki” effect on recent pro-ISIS networks


In 2014, the FBI received a call of concern from a man notifying them that his colleague Omar Mateen had been watching videos of Yemeni-American cleric Anwar al-Awlaki, an al-Qaida propagandist. The man was worried as his friend found the content to be “very powerful”. A few months ago, Mateen walked into an Orlando nightclub, shot, and killed 49 people. This begs the question-- who is this Awlaki and how exactly does he influence people like Omar Mateen to commit acts of violence? According to Seamus Hughes from George Washington University’s Program on Extremism, “..ISIS fighters quote him online and offline..people become radicalized before they become violent. And Awlaki provides the mood music.”

Background: Anwar al-Awlaki
Anwar al-Awlaki, was born in Las Cruces, New Mexico and spent his early years in the United States before returning to Yemen with his father where he spent his teenage years. He came back to the U.S. to pursue an engineering degree from Colorado State University and quickly became a leader at a local mosque in Colorado. In 2011, he was killed by an American drone strike following his involvement in a 2009 shooting that killed 12 U.S. soldiers in Fort Hood, Texas. Awlaki had an ability to radicalize English-speaking individuals online through his widespread videos and lectures that argued for Jihad’s required role in Islam as is fasting, praying, and other religious duties.  His teachings have inspired violent extremism across the world. Roshona Choudhry, a British woman convicted of attempted murder of a British parliamentarian, for example,  claims her ideas for a path of vengeance came from “listening to lectures by Anwar al-Awlaki” she found on the internet.

Twitter as a platform for Connection
A recent infographic by Fifth Tribe, a DC digital agency indicates that 90% of ISIS social media is conducted via twitter with over 70,000 Twitter accounts globally. With over 20,000 foreign fighters now in Iraq and Syria, it is not surprising why one might consider Twitter the “cornerstone of the group’s digital strategy”. To this extent, twitter is a useful platform in understanding Awlaki’s influence among Pro-ISIS users using the data set below.  

Data set:

  1. Kaggle: In my search for analysis on Anwar al-Awlaki’s influence on twitter,I came across a comprehensive data set of over 17,000 Pro-ISIS tweets compiled since the November 2015 Paris attacks (until May 2016). I signed up on the website and downloaded the .xlsx document corresponding to my research.
  2. Anwar al-Awlaki’s ties to extremists (Counter Extremism Project)-- this is a list of 1. prosecuted homegrown radicals with ties to Awlaki and 2. European extremists. I will cross reference this list with my twitter data.

Primary Research Question: How does Anwar al-Awlaki connect and influence recent Pro-ISIS english-language twitter supporters across various popularities, locations, and roles?

  • Sub-question (Popularity)  Does Awlaki connect the more popular (level 3, 4) to the least popular nodes (level 1, 2)?
    • This is important in analyzing themes of recruitment between the most popular and least popular nodes.
    • Network Measures: Newman Girvan (Look at closely connected communities of nodes between the two various levels). I will also look at directed centrality measures of in and out degrees for this.
  • Sub-question (Location/Reference): What are the connections and themes between in the way Anwar al-Awlaki is referenced (example: Quote vs. Lecture) across locations?
    • This is important in analyzing the themes behind his influence geographically and what methods and recruitment materials resonate with certain regions most.
    • Network Measures: Clique analysis to analyze completely connected subgroups by location. Factions: to analyze parts of the network that are more tightly connected than others.
  • Sub-question 3: (Role/Reference )  Who are the brokers between various roles for users (example: Is there a connection between intellectuals and Mujahideen) and the way Awlaki is referenced (lectures/quotes).
    • This is important in analyzing who the bridges and brokers are within the network by the role they play and the way they refer to Awlaki.
    • Network Measures: In and out Eigenvector, and betweenness  in both un-directed and directed network.
  • Location

  • 1. US Based
  • 2. Middle East
  • 3. Europe
  • 4. “Worldwide” /NA

  • Popularity (Number of followers)

  • Level 1: 0-300
  • Level 2: 300-500
  • Level 3: 500+
  • Level 4: 1000+

  • Role: What type of roles do Pro-ISIS users based on Infographic report identifying different types of roles of pro-ISIS twitter:
  • “Reporters” (convey breaking news)
  • “Re-connectors” (users that retweet usernames of accounts  
  • “Intellectuals” ( users that use intellectual justification for what IS is doing)
  • “Fanboys” (users that post messages supporting or celebrating IS)
  • “Recruiters” ((individuals that direct potential recruits to accounts or private messages)
  • “Mujahideen” (individuals fighting on the ground)

  • Reference: How do Pro-ISIS group refer to Anwar al-Awlaki in their tweets?  My initial analysis indicates that pro-ISIS Twitter users in the dataset refer to Anwar Al-Awlaki in the following ways. This might change upon further investigation:

  • Lecture (link to a lecture or audio material)
  • Quote- (resonating quote from Awlaki content)
  • Takfiri (Kufr) (reference to ‘infidels’ or hypocrisy)

**This project is crucial in analyzing how radicalization spreads online through a narrowly focused lens on one of the most important and influential modern actors in the spread of radicalization: Anwar al-Awlaki.  As a Fletcher student focused on CVE, I am particularly interested in looking at ways counter-narratives can work to counter-radicalization. To do that, I first have to understand the spread of that very narrative itself.

1 comment:

Christopher Tunnard said...

Excellent idea: the Awlaki network and its continuing influence on ISIS (recruitment, etc.) A potential cornerstone (to use your term) of CVE efforts. Now that you've got your initial ideas down, you need to focus your efforts on coming up with a more meaningful question. Instead of "How does Awlaki connect...?" consider something like "Can an analysis of Tweets...demonstrate the creation of enduring networks...?" This idea is that SNA of the Twitter nests over time can show what "sticks" and what doesn't, in terms of hashtags and central users. You also need to think more about what SNA measures would be helpful and meaningful.

This is a fascinating and doable project, and it looks like you've got a good data set to start with (Kaggle.) I encourage you to do this, or at least dome version of it. When you're ready, let me know.