To analyze in greater detail the nature of the 66 important jihadist accounts listed by Ahmad 'Abdallah, we extracted the profile data for each account through the Twitter API (see figure 1).¹⁵ Two elements that indicate the nature of these accounts are the common languages and the locations of users. This data shows that the majority of accounts were set up in Arabic (56 percent), although a significant number were set up in English (41 percent), with French making up the remainder (3 percent).¹⁶ This is despite the list's being composed specifically of Arabic-language accounts, which suggests that some of the 66 accounts may be bilingual. 

A second step in understanding this group is to have a sense of where they are. This task is made slightly harder by the fact that few of these users enter meaningful locations in the "location" field on their Twitter profile, and fewer still enable geotagging of Twitter content.¹⁷ Fortunately for our study, however, a surprisingly high number of people set their phone's clock to their current time zone, probably because this allows the time stamps on the tweets they view to make sense. The phone system is set up to define most time zones by a location, such as a major city, which gives an approximate indication of the longitudinal zone in which a user is located (see figure 2).

Humans are creatures of habit. We like the clock to show the correct time, whether it's where we actually are or based on a location with which we mentally associate ourselves. For example, following the 2009 presidential election in Iran, there was a brief campaign for Twitter users to show support for the protesters by changing their phones' location to Tehran, perhaps only to confuse Iranian authorities. This strategy had more than a few problems, however, as writer/ researcher Evgeny Morozov pointed out at the time.¹⁸ One of the more notable problems was the failure of the less-savvy Twitter users to change the time zone as well as the location. Another problem was the tendency of slacktivists¹⁹ to use different tags, such as #helpiranelection, from those used by the protesters or "digital insiders" (e.g., #GR88, #Neda, #Sohrab). As a result, particularly engaged users could be identified by authorities because their interactions on Twitter were predominantly characterized by a series of locally resonant hashtags.²⁰

Casablanca, Morocco, was the most common location that account holders on 'Abdallah's list used to signify their time zone. While it does not mean they were physically in that or any other specific city, it does indicate the area of the world they were likely to be in, particularly given that over 40 percent of the account holders used just three of the more than 250 possible locations available on Twitter profiles (see figure 2).²¹ This suggests that users are most likely to be in Europe, North Africa, and the Middle East, which is consistent with much of the content of the accounts. 

When Twitter users look up the 66 accounts, one of the first elements they tend to examine is the number of followers an account has. While this is not a direct measure of influence, it does indicate that users have heard of the account, and that they may be interested in seeing more of its specific content. 

Of the 66 important jihadist accounts identified by 'Abdallah, @mujtahidd is exceptional, with over one million followers as of 20 August 2013 (see figure 3). The next most followed were @IsraeliAffairs, with around 180,000 followers, and @1400year, with around 30,000 followers at that time. Six months later, in February 2014, @mujtahidd had over 1.3 million followers and @IsraeliAffairs had over 200,000, while the number of followers for @1400year jumped to over 100,000. This data shows that some of these accounts are able to reach and engage a relatively large Twitter community. These three most significant users are discussed in greater detail in the following section.

Figure 2: Time Zone Locations

Figure 3: Number of Followers of the 66

The Three Most-Followed of the Recommended 66 Users of Influence

1. @Mujtahidd 

Mujtahid is an Islamic term of jurisprudence, a "legist formulating independent decisions in legal or theological matters, based on the interpretation and application of the four usul," as defined in Hans Wehr's Arabic-English Dictionary.²² According to Wehr, mujtahid can also simply mean "industrious, diligent." The @Mujtahidd account holder's bio on Twitter consists merely of two Arabic names, Harith and Hummam ("Lion" and "Cultivator," respectively), and his e-mail address, mujtahidmail@gmail.com. These two names also serve as a code related to a saying of the Prophet Muhammad, who declared that these two names were the most dear ones to God and to himself, second only to the names 'Abdallah ("Servant of God") and 'Abd al-Rahman ("Servant of the Merciful"). For his location, the account holder simply entered "the world." Jihadist Twitter users claim that the account holder behind @mujtahidd is a "known whistleblower inside the #Saudi government," according to Al Battar Media, a media outlet that merged with the SSI media front and became the forum's official media arm on 8 January 2014.²³ The "whistleblower" designation serves to enhance the account holder's cachet as an insider working against an apostate regime on behalf of God and the jihadists' cause.

2.@IsraeliAffairs 

@IsraeliAffairs had about 180,000 followers as of 20 August 2013. The account holder writes in his online bio, "I am Muslim, my citizenship is Arab, I work on behalf of my country which is every span of hand on earth; raising on it the adhan [the call to prayer]! [I am a] diplomat, translator, researcher in Israeli affairs." 

3. @1400year 

@1400year, the third most-influential account, has about 85,000 followers. The Arabic name of this account is gharib fi wadanihi—"the stranger in his own country." The sentiment of gharib is a reference to a frame of mind: the "true believer" considers himself to be something of a foreigner in this world and associates himself with the earliest Muslims, who had also been perceived as strangers in their own historical time and milieu. The eagle next to the ISIS flag on the account's Twitter page is a clear allusion to this idea of jihadists as "strangers" (ghuraba' ). "Stranger" or "estranged" is used here in the context of the Israeli occupation of Palestine, which is considered to be the "first catastrophe" (nakba) in a series of dramatic events that have affected Islamic countries or territories since the mid-twentieth century. Jihadists are convinced that there is a global "conspiracy against Islam," hence the British grab of Palestinian land followed by the declaration of the state of Israel (1948), and subsequent epochal conflicts including the occupations of Afghanistan (1979/2001), the U.S. invasion of Iraq (2003), and the Arab Spring conflicts after 2011, have entrenched this notion in jihadist communities both on- and offline. 

When the data for @1400year was captured in August 2013, the bio for the account holder stated, "The man in the picture is Rachid Nekkaz, a French millionaire of Algerian origin, who opposes France's ban of the niqab. He said to the Muslimat of France to wear the niqab and I will pay the fine, I am honored by placing his picture [on my account]."²⁴ The account holder's updated bio, however, states, "The demise of Israel may be preceded by the demise of [Arab] regimes that made a living on the expense of their own people, laughing at them, destroying the societies."²⁵ 

@1400year also has a YouTube account linked to it with 1,830 subscribers and 437,243 views. Having such links across platforms allows users to more effectively create their zeitgeist. This is similar to the way jihadist groups such as Jabhat al-Nusra are using Twitter to disseminate links to Syrian battlefield videos posted on YouTube.²⁶ 

Of the remaining accounts, 31 of the 66 accounts listed have between 5,000 and 100,000 followers, as shown in figure 4.

The mean number of followers for this group of accounts is 28,220, but this is heavily influenced by the three accounts with the greatest number of followers. The median number of followers is much lower, at 5,377. This gives an indication that important jihadist accounts, or at least those thought important in 'Abdallah's Twitter guide, have the potential to reach a relatively large number of users, especially when compared, for example, to the U.S. State Department's Digital Outreach Team and its Arabic Twitter account, @DSDOTAR, which had a little under 900 followers as of early 2014. 

Figure 4: The Next 31 Most-Followed Accounts

Do These Accounts All Speak to the Same Followers? 

This overview of the 66 "important jihadist accounts" can give the appearance that collectively, the accounts are reaching over 1.8 million users. However, as noted earlier, @mujtahidd alone is followed by over 1.1 million users, and is such an outlier that we have excluded it from the subsequent analysis. Even so, the real number following all of the remaining 65 accounts is much less than 700,000, because a number of users follow more than one account. By building a network representation of the users who follow the important jihadist accounts, we found that the network following one or more of these accounts (excluding @mujtahidd) was a little over 370,000 users, which gives a more accurate picture of the reach these groups collectively achieve. 

To break it down further, many users follow one or two accounts, as we might expect from an online social environment, while a very few follow several of the listed accounts. Figure 5 illustrates a close approximation of a statistical power law curve.²⁷ 

Of those users who follow only one of the 65 (minus @mujtahidd) important jihadist accounts, 34 percent follow more than one. Of the users who follow more than one of these accounts, however, 45 percent follow only two accounts. These can be thought of as casual followers. At the other end of the scale, there are 504 users who follow 40 or more of the listed accounts and 109 users who follow 50 or more. These followers clearly are more engaged than most. 

The fact that a user is particularly engaged in following the accounts deemed important in 'Abdallah's Twitter guide does not necessarily indicate any political affiliation—not least because of the number of CT scholars and professionals who actively follow such accounts. It is, however, instructive to consider the aggregated traits of this "highly engaged" group of 504 users who follow 40 or more important jihadist accounts. 

Unsurprisingly, given the dominant languages used by the 66 listed accounts, Arabic, English, and French are the languages that appear most frequently in the Twitter profiles of these highly engaged followers (as shown in figure 6). In addition, there is a small number of users who post to Twitter in other languages, such as Indonesian, Spanish, Dutch, and German. These users are likely to be multilingual, given that the content they disseminate is in Arabic, and they may act as bridges by connecting the core content in Arabic to wider communities that do not speak Arabic. 

From the aggregated profile data, a similar question can be asked about where in the world these highly engaged users appear to be. 

The findings from the location data (see figure 7) highlight the cultural importance of appearing (at least) to be located in or near the Arabian Peninsula: almost 20 percent of highly engaged followers use Baghdad to set their time zone. Collectively, the followers' data shows a focus on the Middle East and Europe similar to the 66 important jihadist accounts. 

This section has shown that the 66 accounts listed by 'Abdallah, as we might expect, tend to tweet in Arabic. They are collectively followed by a network of around 370,000 people (if @mujtahidd is excluded), but most of these are casual observers who follow only one or two of the accounts. There are, however, 500 to 1,000 more-engaged followers. These active followers tend to be Arabic speaking, have relatively few followers themselves, and appear to have a greater tendency than less-engaged users to identify with the Arabian Peninsula region, and Baghdad in particular. 

Figure 5: Frequency of Users Following Important Accounts

Figure 6: Languages Used by Highly Engaged Followers

Figure 7: Time Zone Locations of Highly Engaged Followers

Relational Dynamics between the 66 Important Jihadist Accounts 

An analysis of the relational dynamics between the 66 accounts on Ahmad 'Abdallah's list provides insight into the way these accounts relate to each other, and the relative importance of each account to the others on the list. These relationships are important because they influence the way individuals search for information, what they find, and the behaviors they adopt.²⁸ 

When we reconstructed the network of follower–following relationships, we found that @JbhatALnusra, @WaleedGaj2002, and @AsadAljehad2 are most frequently followed by the other accounts on the list. Conversely, @SaveArakan4, @Mhaajrr, @housse_100, and @alassra2012 appear on the list but few of the other accounts follow them. The ranking for each account (see table 1) is based on how frequently each account is followed by the others. This ranking is also comparable to the ranking produced by the eigenvector calculations for the network (see figure 8).

The network as a whole represents 958 relationships among the 66 important accounts, with a network diameter of 5 (the distance between the furthest two nodes calculated on the directed graph). The network density is 0.2 on the directed graph, meaning around 20 percent of the connections that could exist actually do exist. (A score of 1 would represent a complete graph, where all connections would exist.) 

Table 1: Rankings
 

Figure 8: Eigenvector Calculations

Key Nodes in the Network 

We found, perhaps not surprisingly, that the majority of the connected accounts are hardcore jihadist media activists, of which the most frequently followed is Jabhat al-Nusra. In this section, we look at some of the high-profile accounts whose owners are clearly advocating the full extent of violent jihadist ideology on a global level. 

@WaleedGaj2002 

This account has over 45,000 followers, while following 387. It apparently belongs to Walid (Muhammad) al-Hajj, a "former Guantanamo detainee and eyewitness of the Qila-e-Jangji massacre in Afghanistan,"²⁹ who is from Sudan. According to the WikiLeaks Gitmo Project, as reported by the New York Times, he was released from the Guantánamo Bay prison in 2008.³⁰ Al-Hajj's tweets indicate that he still supports the mainstream al Qaeda jihadists: he praises Osama bin Laden while being cherished by other Twitter users for having known the shaykh personally, and he is also active within the mainstream media.³¹

@Strategyaffairs 

The account holder behind @Strategyaffairs is a prolific and quite industrious jihadist media activist, with over 100,000 followers and just over 4,000 tweets as of August 2013. He is also active in most classical jihadist forums as 'Abdallah bin Muhammad. Occasionally he has tweeted statements by the Yemeni group al Qaeda in the Arabian Peninsula (AQAP) prior to the "official" broadcast within the forums, and provided context for as well as posted AQAP's statements, both within the forums and on Twitter. Other SSI members, for instance, used his Twitter input to further the AQAP statements within the forum, thus raising the status of @Strategyaffairs in general.

@EYADQUNAIBI 

Dr. Iyad Qunaybi is something of a rising star within the radical on- and offline scenes. He is active on all social media outlets, and his videotaped speeches are also transcribed and published within the jihadist forums by media outlets such as the previously mentioned @fursanbalagh. His sermons are presented in Arabic and sometimes subtitled in English. Qunaybi rose to fame within this subculture in 2011–2012, during the uprisings in Egypt, but his focus has been on Syria since the outbreak and spread of violence there. His standing was boosted when his posts began to be included on Abu Muhammad al-Maqdisi's database, under the name Iyad al-Qunaybi.³² His input is valued, retweeted, and disseminated within the jihadist forums and on Facebook, and is further published on sites such as JustPaste.it, where content is easily shared and almost impossible to track. 

@as_ansar 

This is the official Twitter account of the bilingual Shabakat al-Ansar al- Mujahideen, a tier-one jihadist online forum. The main forum is in Arabic, with a sister forum in English. Both forums are sometimes taken down, but they usually resume working after some time. The instability of the main forum is a good example of why Twitter has become a tangible alternative for the media-driven jihadists—the Twitter accounts remain alive and very active with their over 26,000 followers, untouched by any disruptions to the online forums. The YouTube link shown in the @as_ansar screenshot at left is an "invitation to Muslims to visit the forum" and instructs visitors on the use of the Tor Project privacy app to conceal their identity online.³³ 

@Al_nukhba 

This is one of the pioneer jihadist Twitter accounts, advertising its founders' passion for social media from as early as 2009. The name nukhbat al-I'lami al-jihadi, the "Jihadist Media Elite," may stem from Abu Mus'ab al-Suri's July 2005 "Message to the British and European Peoples and Governments regarding the Explosions in London," in which he described the internet as the most important medium through which to propagate and spread the jihadists' demands, and ideology in general.³⁴ He called on "the jihadist elite" residing in Europe to partake in this venture. 

The @Al_nukhba account could be described as a jihadist media hub. Its members, active on the forums for years, are highly committed and regularly provide transcriptions of jihadist media productions. This is naturally very helpful for any analyst, but it is also quite a service for the jihadist audience. The speeches of main leaders and ideologues, as well as major video productions of such outlets as Sahab and al-Malahem, are transcribed and can be conveniently downloaded as a PDF or Microsoft Word document. On the group's main website, Nokbah.com, data collections and videos can be downloaded and also searched. It is a well-built and -maintained data warehouse for extremist content, which is uploaded and disseminated primarily via the classical forums. @Al_nukhba has about 6,000 followers, had tweeted over 500 times by early 2014, and was following no one. 

The "betweenness" calculation (see table 2) highlights those users through which the shortest paths across the network most frequently pass. These users are often found near the center of the network image. From the perspective of betweenness, in addition to some of the users mentioned previously, these accounts also appear to play an important role as bridges between different elements of the network. To have a high betweenness score, users have to both follow and be followed by other users.

Table 2: "Betweenness" Centrality

@Caucasusaffairs

@Caucasusaffairs has a significantly higher betweenness ranking compared to some of the more-followed accounts: @Caucacusaffairs is both widely followed and an active follower. As the name of this account implies, @Caucasusaffairs covers everything related to the Caucasus region and the organization called the Caucasus Emirate, with a focus on Chechnya. The tweets published on the account are mainly in Arabic, but some are also in Russian. @Caucasusaffairs has over 44,000 followers while following over 400 users itself, and has posted about 8,000 tweets. Such volume makes this account a valuable asset in addition to the main forums and their relevant subsections. It mainly retweets the Arabic-language media outlet Echo of Caucasus (through @RevOfIslam), which has been one of the main media hubs in Arabic for many years. The image of a four-fingered black hand on a yellow background (see figure 9), which is widely used on Muslim Brotherhood sites, is a symbol of the massacre that took place at Egypt's Rabaa al-Adawiya mosque on 14 August 2013, when Egyptian security forces destroyed an encampment of supporters of ousted President Mohammed Morsi. This symbol "has emerged in the Middle East, online and offline, to remember the crackdown of the Rabaa al-Adawiya protest camp whereby many pro-Mursi citizens were killed."³⁵

The use of the Rabaa al-Adawiya icon in the Chechen context is yet another attempt by the media-savvy jihadists to globalize their agenda, and serves among them as proof for the repeated claim that there is a "war against Islam." This icon has been further "jihadized" by adding the movement's black banner, as shown in figure 9.³⁶ 

This analysis of the relational dynamics between Ahmad 'Abdallah's 66 "important jihadist" accounts through their following–follower relationships shows that the accounts form a relatively dense network, which has two results. First, it creates mutually reinforcing clusters of information, which can crowd out other perspectives,³⁷ and thus contributes to the development of a zeitgeist, or a new electronic propaganda frontier, as was discussed in relation to the activity of Jabhat al-Nusra.³⁸

Second, the density of the network tends to protect it against basic disruption strategies such as the removal or suspension of individual accounts. As Paul Baran's work On Distributed Communications demonstrated, only a small level of redundancy is required to build a communication system that can withstand heavy enemy attacks.³⁹ Although Baran's work was done in the pre-internet context of the 1960s, and focused particularly on media problems faced in the 1970s, the insight provided by the study also relates to online activity and the need for more complex strategies to disrupt dense communication networks.

Figure 9: Rabaa al-Adawiya Protest Symbol

Information Flow

This final section assesses the flow of information in a network that includes the 66 important jihadist accounts. Tracking the spread of this information provides a useful insight into the way relationships can both influence the spread of information and allow the identification of key actors or influential users within the jihadist Twitter phenomenon. For example, a previous article by the authors on the Syrian jihadist group Jabhat al-Nusra demonstrated how jihadist groups are using Twitter to post links to video content shot in Syria.⁴⁰

Some of the data from the first sections of this article was posted on the blogJihadica in August 2013, along with a brief analysis of the 66 important jihadist accounts.⁴¹ In this article, we are making use of data about how Twitter users communicating in Arabic responded to the information we posted on the blog. Following the post, which is in English, requests appeared on Twitter for a translation to be posted in Arabic. This translation subsequently appeared on JustPaste.it,⁴² and news spread rapidly on Twitter using the hashtag #يداهج_باسح_ 66

_رطخأ_ةيكيرمأ_ةسارد

The majority of the tweets containing  _ةسارديداهج_باسح_ 66 _رطخأ_ةيكي# ‬appeared between 8 and 10 September 2013, after which there was little further activity. At its peak, the hashtag was used over 800 times in 15 minutes, equivalent to 53 tweets a minute (see figure 10).

The graph in figure 10 shows the volume of tweets containing the hashtag for each 15-minute period. This indicates that the news traveled fast, reaching most of the users it would ever reach within the first day, and highlights the speed of information dissemination within this network. A few users were particularly important in spreading the news. Figure 11 shows the users most frequently retweeted within our data.

In addition to the volume of retweets, time is also a factor, because users were retweeted or mentioned at different times. Known as an engagement profile, the graph in figure 12 shows the time period in which users were particularly important.

Figure 12 also highlights that prominent users are retweeted rapidly; if this observation held up in numerous other contexts, this would indicate that information is traveling across such networks faster than current case-by-case counterstrategies can respond. Among the 66 jihadist accounts in this study, @Tuohed and @almohajermuslm had the greatest levels of retweets and mentions early on 9 September 2013. Twelve hours later, @xxggxx2 became prominent, and nearly 12 hours after that, retweets of @albttarm peaked. This hints at the features of the networkthat give it greater resilience. Information can be shared rapidly, as shown by the initial spike in retweets. The engagement profiles also show, however, that different users are able to continue the conversation, which highlights the fact that prominent users fulfill different roles in the network.

To gain greater insight into the way information travels and the role of prominent users in the network, key actor analysis (see figure 13) can differentiate those users who are important for reaching specific communities from those users who are at the core of the network. If there are a number of users who operate within the core, or vanguard, and others who act as conduits for content to a wider audience, this further enhances the network's resilience by allowing it to withstand disruptions caused by the occasional suspension of accounts that apparently breach Twitter's terms of service.

The position of key actors on the scatter plot in figure 13 is based on two network metrics: betweenness centrality and PageRank. Betweenness centrality refers to how often a node lies on the shortest path between any two nodes in the network.⁴³ A node with high betweenness centrality has the potential to influence the spread of information, by facilitating, hindering, or even altering the communication between others. PageRank is an objective measure of the importance of a node in a network—a specific citation graph. As Google cofounders Sergey Brin and Larry Page describe this method, the result "corresponds well with people's subjective idea of importance":⁴⁴

• Users in the bottom left of the graph tend to have no particular role and can be thought of as general users, although they may have high value to a very specific, often very small, group.
• Those in the top left of the graph tend to be in the core (or one of the cores) of the network. This indicates they are often the ones most invested in the network and have access to privileged information, but they rely on others to disseminate the original content they tweet to a wider audience.
• Those in the bottom right of the graph fulfill the role of bridging between the core content producers and a specific community. The value of this role often comes from tailoring information to a specific audience, which makes these users more valuable to that group but less important to everyone else.
•  Users in the top right are rare. They serve a dual function: they have the same trusted status as those in the top left of the graph, and they also fulfill the same "bridge" role as users in the bottom right of the graph, reaching areas of the network that others do not serve.  

​​The network map (see figure 14) can show which users communicated with each other, which reached the same communities, and which were a bridge to specific communities.⁴⁵ The colors indicate the different interconnected communities; note, in particular, the green cluster of interconnected users in the middle of the network. This highlights the likely resilience of the core of the community on Twitter, in addition to the resilience created by the multiplatform zeitgeist identified in our earlier study of Jabhat al-Nusra.46 There are significant levels of redundancy in the connections at the core of the network, which allow communication to continue even when some accounts are suspended.

Figure 10: Arabic Speakers' Response to Jihadica
Article 8-10 September 2013

Figure 11: Most Retweeted Users

Figure 12: Response to the 66

Figure 13: Key Actor Analysis

Figure 14: Network Map

Conclusion

This analysis demonstrates that jihadist groups have used the opportunity created by the proliferation of social media platforms to develop a network of Twitter accounts that gives them a persistent as well as ideologically cohesive presence for jihadist propaganda online. The case studies presented here indicate that the level of interconnection between core members of the network has achieved a high degree of resilience against disruption from the suspension of individual accounts. If this finding can be repeated across a number of other examples, it will have wide-ranging consequences for counterstrategies, which will need to shift from countering individuals to disrupting cohesion across a network if they are to be effective.