Sander van Haren
Mitch van Helvert
May Sundvall AndersenSerkan Yıldızeli Special thanks to our Italian analysts/designers
Carlo De Gaetano
Tommaso Renzini[ Back to top]
In the Netherlands there is an emerging division between anti-Islamism on the one hand opposed to the more radicalizing Muslims. The Dutch political climate has been feeding the opposition between Muslims and non-Muslims. Politicians such as Geert Wilders has come with polarizing statements that send the message that the immigrants are no longer welcome in the Netherlands. The murder of Theo van Gogh, who was publicly opposing the Turkish and Moroccan immigrants, and the increasing news of attacks by fundamental muslims in the western world (think 9-11) led to a wave of anti-Islamism in the Dutch society, increased the gap between the opposing sides.
There is a growing concern about the numbers of Dutch citizens heading to Syria to join Jihadist trainings camps in Syria. The latest estimate indicates that some hundred Dutch muslim extremists are involved in the Syrian civil war, according to numbers of state agency Nationaal Coördinator Terrorismebestrijding en Veiligheid (NCTV). NCTV fears the fundamentalists [that go to Jihadist trainings camps in Syria] later return to the Netherlands with serious fighting skills and Jihadist thoughts. This can form a serious thread to national security.
Muslims in Syria get in contact with muslims in the Netherlands through social media such as Facebook. Some try to recruit people to join the Jihadist movement. National security tries to track these social media groups in order to prevent Dutch citizens traveling to the trainingcamps.
This essay is the result of a project of the winter school at the University of Amsterdam. Trying to understand how this tracking works and how it could be approved in this essay we track five Dutch Facebook pages that have an extremistic Islamic appearance. We have asked ourselves the question: 'How are the main actors in the Dutch Jihadist Facebook network identified through a network- and interaction analysis of the top pages, and what are the remarkable characteristics of these pages and key users?
By conducting a network analysis on the five Facebook pages we reveal and visualize parts of the Dutch Jihad network by locating key actors. With this we will paint a picture of how Facebook serves as a source of information, a place to discuss and recruit people for Islam extremism.[ Back to top]
We start out by explaining Jihad, what the term means in general and what we mean when reffering to it. After this we continue with the theoretical framework on how and why extremist could use the web and Facebook. Then we discuss the methodology by a brief history of social network analysis , its key concepts and advantages and disadvantages before we do explain our approach and research structure. Having explained this, we go on to our findings, visualized through graphs and images, before we continue to a discussion of our findings. In the end we summarize our research through a conclusion.[ Back to top]
In news media however, Jihad is most often used to define the terrorist attacks that are done in the name of the Islam. In means of the holy war in this definition of Jihad, it is a religious duty to spread Islam with force if necessary. Because this it the term most often used in news media as well as on the Facebookpages it is the last definition we mean when using it in this essay. Furthermore, for the purpose of understanding picture analysis it is necessary to explain the uses of the banner that has been used for Jihadist to express themselves from at least 2001. The flag has been called the banner, the banner of the eagles or the Black Standard. The banner can be formed differently, but seeing them in the Facebook pages means that we are dealing with extreme Islamic and Jihadist actors. Below you can see two of the most used pictures of the banner.
Extremist groups have been organizing on the internet for some time. They have their own websites in an area called the Dark Web, a term by Mike Bergman. For this research the Dark Web is a place on the World Wide Web that is being used for extremist or terrorist activities. The possibility to reach a large group of people worldwide in a very short time note makes the internet a perfect place to organize and recruit new members. Chen et al. (2007) argue that with traditional media (newspaper, radio and television) it is difficult for extremist groups to spread their message and ideology. Since they usually do not meet the editorial selection criteria their messages are not accepted on the these media. The internet as a space without editorial selections they can spread their message. Researchers at the Institute for Security Technology Studies (ISTS) have defined five categories in which terrorist groups make use of the web: propaganda, recruitment and training, fundraising, communications and targeting. These five categories are all focused on member activity and enhancing their information operations (Chen et al: 2007). By posting one-sided information groups can propagade online. With propaganda they try to convert and recruit people. Next the training starts. Guides about weapons and tactics can be uploaded and be read by members. Video games can be altered in such a way that they more or less recreate real-life situations so that combat tactics can be trained and studied. As money organization and material cost money donation pools are set up online and financial contributions are asked of members. Illegal means of gathering money is also an option for them when using the web (Conway 2006).
Online platforms also provide a space for discussion, meetings or arrangement for meetings. Both locally as globally groups and people can stay in contact, keep each other up to date about the current situation and exchange new information. Showing the results of their actions or by showing that they abducted people and are demanding ransom for the release (Conway 2006). By using the internet for these means extremist also get publicity in the traditional media, because they cover the stories found on the internet and show videos that extremist have uploaded on the web on the television. The various platforms on the internet make it easier for these extremist groups to target certain demographic groups, who might have been more difficult to target before. They can, for example, more easily target women online as this demographic group is more active on Facebook (Torok 2011). Targeting potential victims or locations for attack are done the same way. Most of the categories are connected and enforce each other. If their propaganda is strong and extremist target the right groups are more likely to get more recruits. Recruiters are not only needed for actions taken but it can raise fundraising as well spread the message of a large organisation. This might intimidate opposing parties and attrackt ally parties. This can all be done by creating a good communication network and be able to reach everyone. However, using sites that extremist build or host themselves can be quite exposed to unwanted eyes and ears, for example the United States or other Western countries.
With the emergence of social media and websites like Facebook, extremist groups have moved to use these more often. A downside of creating and using own websites is that they are vulnerable to DDoS attacks from hackers, causing the server to overload and it brings down the websites. The United States and their allies also target the websites of these extremist movements and monitor them (Waskiewicz 2012). When using Facebook extremist can protect themselves from DDoS attacks more easily. It is much harder for hackers to launch DDoS against Facebook than against their own websites, since they will not only harm the extremist groups, but the hackers will also harm commercial companies and other users. Facebook is providing a more protected environment against hackers. Facebook provides them a free environment for them to use, since the can freely create open and closed groups. Open groups are an ideal place for them to attract those interested to more specific pages and to create a network. Closed groups can be used for more extreme messaging and activities. The fact that the creators of the pages are the one that can decide whether or not to accept new members make these pages a more secure place. Users can also request to join but the admin of the group has to approve the request. This allows them to first screen the person who wants to join the group, before allowing them access. The content members post in the group or on their own profiles can be monitored by other members of the group. Making it easier for them to keep members in check and make sure they do not leak information to the open. On Facebook, groups can like and follow each other to stay connected and create a network. Making it easier for them to coordinate with each other and so execute plans in the real world.
Extremist use a form of friend of a friend structure. Person A influences person B who on his turn influences person C. Even though there is no connection between person A and C, when the influence between all three is strong enough A also influences C. This kind of network structure is what these kind of groups are trying to realise (Waskiewicz 2012). They have certain tactics to not just only recruit males, but also females and Facebook is a place for them to find many women. Sue Mahan and Pamela Griset have found four categories of the roles that women fulfil within the organisation: sympathisers, spies, warriors and dominant leaders (Torok 2011). The way extremist recruit women is a long process. They are welcomed and swarmed by the members and most of them are willing to interact with them. Members are open-minded and slowly expose women to their ideas. Extremist post content that shows oppression against Muslims and so expose women to their propaganda. Members are also being used to track and follow military personnel from Western countries. Soldiers have Facebook profiles and on it they might share information about themselves, where they are going on holiday and other personal information (Weimann 2010). This makes Facebook a good place for extremist to gather intel on their enemies and where they are going to. It might also be easier for women to befriend soldiers, so that women can monitor their Facebook profiles.[ Back to top]
There are key concepts to the heart of network analysis that are fundamental to the discussion of social networks (Wasserman 17). We will now briefly explain the concepts.
Actor: As mentioned earlier social network analysis is concerned with understanding the stucture of social orgaization by studying the linkage between social entities and the implications of this linkages. The social entities are referred to as actors ( Wasserman 17). Actors can be discrete individuals, corporate or collective social units . It is important to mention that the term actor does not imply that these entities have the will or ability to act (Wasserman17). A collection of actors is called one mode networks (Wasserman 17).
Triad: The term triad refers to relationships among larger subsets of actors (Wasserman 19). As the term implies triad refers to the study of three actors and the (possible) tie(s) among them (Wasserman 19). When studying triads of particualar interest is wheater the triad is transitive (if actor 1 likes actor 2 and actor 2 in return likes actor 3, then actor 1 will also like actor 3) or balanced (if actor 1 and 2 likes each other, then 1 and 2 should be similar in their evaluation of a third actor, 3, and if 1 and 2 dislike each other, then they should differ in their evaluation of a third actor, 3) (Wasserman 19).
Subgroup: Dyads are pairs of actors and associated ties, triads are triples of actors and associated ties (Wasserman 19). Locating and studying subgroups using specific criteria is a important concern of social network analysis (Wasserman 19). We can define a subgroup of actors as any subset of actors, and all ties among them (Wasserman 19).
Group: Network analysis does not only consist of location dyads, triads and subgroups. Network analysis is largely about visualizing a model of the relationships among systems of actors (Wasserman 19). Systems concists of ties among members of some group (Wasserman 19). A group is the collection of all actors on which ties are to be measured (wasserman 19). One must be able to argue by theoretical, empirical, or conceptual criteria that the actors in the group belong together in a more or less bounded set (19)
We have now explained the key concepts of network analysis with its advantages and disadvantages. All the mentioned key concepts create a working vocabulary for discussing social networks and social network data (Wasserman 20). Since we will be focusing on the specifically the Jihadist Facebook networks with relation to the Dutch situation, it is of great importance to outline a clear methodology. The strategy of the research will then be related to answering the main question. The journey to posing an answer to the question will be made possible through the use of different set of tools and approaches.
There are both advantages and disadvantages when using network analysis. One advantage, that is quite obvious, in doing network analysis is that conclusions can be drawn about a large quantity of connections that otherwise would seem irrelevant or too complicated. Very complex and incomprehensible amounts of connections can be made clear and structured.
What we should be careful of with doing network analysis is that it is easy to only look at networks to reach the wanted result, thereby overlooking the part of the network that is not involved in the researched network. For example in online network analysis, before people can get online they must have access to a number of (digital) resources that require a certain set of funds and skills. Some people simply do not have these and consequently can not be a part of these networks, or the resulting analysis.[ Back to top]
As a starting point, the research process was initiated with the help of existing insights provided by the Ministry of Security and Justice. These data have also set the basis for the strategic approach on delimiting the corpus. In fact, the data provided by the Ministry outlines the offline Dutch key figures and leaders related to the extremist Jihadist cause.
As derived from the surface data and information by the National Coordinator for Security and Counterterrorism, there have been notable findings. Here it was noticeable that the online Jihadi subculture had been easily accesible and is in fact fairly open and operate within the boundaries of the law. This is of course related to the possibility of monitoring these Facebook pages that, at least to some extent publicly shared information. Consequently, this means that various types of content, connections and interactions on a social media platform like Facebook can be analysed with the use of digital methods tools.
The first tool we used is the Facebook application Netvizz, which is created by Bernhard Rieder (Associate Professor of New Media at the University of Amsterdam). It extracts data from different sections of the Facebook platform, such as the personal profile, groups and pages. The features of Netvizz has enabled us to take on different approaches in applying supportive research methods. We have analysed using the pages and page likes network features of the application. The page data settings were set to get the last 50 posts of the Dutch Jihadist pages that will be specified in the following section. Through the output, we had the possibility to analyse the interaction of indivual users on the page content. Before going into further detail and analysis of the main actors on Facebook, we had visualized the page likes network. The functionality of the likes network of a page is accompanied by the selected page (the "seed"), which then retrieves all the pages that page likes. The application has continued until the specified crawl depth was reached, which had been set to 1. This is done in order to identify certain clusters and networks of possibly Jihadist pages but will also give an overview of the diversity in the page like clusters. The output will be containing a directed network of pages, which will be visualized through Gephi.
Gephi is an open source graph for visualization and manipulation software. For the production of a network, two inputs are needed: a list of the actors composing the network, and a list of the relations (the interactions between actors). As part of a mathematical object, actors will then be called nodes, or vertices, and relations will be denoted as edges.
The relevant pages and their content that will be our corpus will be selected on the basis of the criteria described below. We have outlined a supplemented list through which we will determine the research-relevancy of a Facebook page.
The criteria on which we have based our extreme Islamic pages list is formulated as follows:
- Jihad related visual symbols (flags, training camps, rebels/warriors etc.)
- Symbols related to Al Qaida flags
- Predominantly violent and provocative speech and media
- Images and eulogies of martyrs
- Videos and images from battlescenes and weapons
- Expression of extreme Jihadist discourse
- Glorification of radical religious leaders
- Users and expressions that point to an interest in the Jihad
- Decent amount (+100) of page likes (as an indicator of the value/relevance of the page)
In order to find and analyse the pages with Netvizz, we had generated two undercover Jihadist profiles. Through these profiles, we had liked the pages and consequently made them useable for further analysis with Gephi.
In the following section, the research and analysis will be conducted according to the aforementioned methods and strategies. In addition, a more discursive analysis, assisted by sentiment analyses, will broaden the insights gained through the results.[ Back to top]
The profiles and interests of these two key figures provided a good entry point in identifying the most important Dutch Jihadist pages. After an in-depth analysis of the pages according to our criteria, the following list of the pages have been composed:
- Ahlus-Sunnah Publicaties
- Shaam al-Ghareeba
- Onze Gedetineerden
- Yes We Support The Mujahideen
- Soldier Of Allah NL/BE
Below, you will find the page like networks of the five specified pages. We have appended the graphs of the five page like networks in order to get a total overview of the relevant liked pages. Although the visualized network gives a good overview of the page interests of the Dutch extreme Islamic pages, this does not necessarily mean that every liked page is Jihadi or extremist.
Below you can find the location and size of the invidual network of the previsouly determined pages:
Total (appended) overview of the network:
- Ahlus-Sunnah Publicaties
- Shaam Al-Ghareeba
- Onze Gedetineerden
- Yes We Support The Mujahideen
- Soldier Of Allah NL/BE
The visualizations indicate that, in terms of the inDegree and page likes received from the specified extreme Islamic pages, the Soldier of Allah NL/BE page is the most isolated, whereas Ahlus-Sunnah Publicaties has the largest network.
The size and colors of the nodes are both connected to the value of the inDegree. A higher value will display a larger node with a darker color of green. It is visible that the two mutually most liked pages in the appended network are Shaykh Ahmad Musa Jibril and صفحة الشيخ تركي البنعلي أبي سفيان السلمي (Page Sheikh Turki Al-Benali Abi Sufyan peaceful, Google translate).
Viewing the two most liked pages, which are categorized as Public Figures, with regards to our appended network, we can see a more neutral and educational page of Shaykh Ahmad Musa Jibril, whereas the page of Sheikh Turki Al-Benali Abu Sufyan explicitly features the flag defined as Islamic State of Iraq flag. This implies the orientation of the page towards a Jihadist ideology.
Profiling the top ten most active users of the Facebook pages we started with an overview of profile pictures and banners by using an image scraper (above). In these images we can see a lot of black and yellow in general. Most of the images include some form of Arabic texts and the Black Standard.
Digging deeper in the top ten users of the Facebook pages most of the personal profiles focus on Islamic and Arab content. Pictures of the Black Standard are encountered in nearly all profiles. Also the lion keeps appearing. This seems to be a symbol for the force of the Islamic religion, which is not to be held down or caged. Other pictures as well as the page likes are either Islamic or Arab oriented. Only a small number of users have friends in their list that have non-Arab names. Apart from these profiles almost a third of the profiles is obvious anonymous. These profiles display either no profile picture, or at least not of the person itself. Most of these users seem to use a fake name and have few Facebook friends, if any.Looking at the time the profile pages are set up, the oldest is dated in 2007 and the newest at January this year. Over 90% of the profiles are set up in 2013. The same year all of our Facebook pages are set up as well. Significantly the social media platform Hyves stopped in 2013. Possibily these groups have moved to Facebook pages and the user profiles are set up in order to keep following the groups. In the next section we will examine page specifics and conclude with a general graph of the multiple pages.
type of activity:
Shaam Al-Ghareeba gender: type of activity: location users: Soldier of Allah NL/Belg gender: type of activity: location users: Onze gedetineerden
type of activity:
Yes We Support the Mujahideen
type of activity:
This graphic shows the different pages. In the middle, it shows several nodes. The biggest nodes are the ones connected to multiple pages, with a maximum of 5. Most of these users are also identified in our section about the top contributors (Likes/Comments). The colors show the different pages. Overall, the pages are relatively isolated, with the exception of the core. At the core, the pages come together. This shows that there is a selection of users that are connected to multiple pages. These nodes (users) are interesting to identify because they are connected to multiple pages, and therefore the most active in our scope.
When examining the form of activity on the pages we first noticed that all of the pages are mainly driven by user activity, in the form of comments and following, in case of most pages, with pictures. Only the Ahlus Sunnah-publicaties, the second top-activity is in the form of status updates the page seem to be driven by the administrator. Ahlus Sunnah-publicaties follows a lot of different pages and seems to act as a sort of mediator between related pages. Photos are a popular format in general. Most activity is generated in the form of user reactions, as a reaction to photo posts.
When examining the gender division we encountered some interesting findings. At first, we expected there to be a large majority of males. This expectation can be. Examining the division of male and female participation and following of the pages we discovered that the majority is female and in the cases of page, page, this group are at the same time the most active. This could be explained by what have formerly explained about targeting women mostly online (Torok 2011).
In the last graph, the diversity graph, it is interesting to see that the most users are active in their own group. There is a section of users that is active on multiple pages only a relatively small selection active in all of them. The users that are active on all of the pages could be considered the most interesting. These are the most diverse, connected users in the pages we researched.
In our research, we were able to find the top ten, most active users (likes and comment-wise). However, due to privacy issues we will not publicly publish these names.
This is the top five of most diverse/connected users (they are connected to five pages):
1 U. Nidal
2 U. Yassine
3 U. Mujaheed Shaheed
4 A. El ousrouti5 S. Tawheed [ Back to top]
For further research we recommend an extension of the extreme Islamic network on other social media platforms, or for other countries that connect to this network. As Facebook is an open space where anybody can enter and make (fake) profiles it might be difficult to link entire networks to real life.[ Back to top]
Chen, Hsinchun et al. Uncovering the dark web . A case study of Jihad on the web. Journal of the American Society for Information Science and Technology 59.8 (2008): 13471359.
Conway, M. Terrorism and the Internet: New Media--New Threat. Parliamentary Affairs 59.2 (2006): 283298. Web. 16 Jan. 2014.
Hult, S.J. van. Violent Jihad in the Netherlands. Current trends in the Islamist terrorist thread. General Intelligence and Security Service. Ministry of the Interior and Kingdom Relations. Den Haag 2006.
Moreno, J.L. Who Shall Survive? New York: Beacon House, 1953.
Morgan, Diane. Essential islam. A comprehensive guide to belief and practice. Santa Barbara: Greenwood publishing group, 2010.
Qin, Jialun et al. Analyzing Terror Campaigns on the Internet: Technical Sophistication, Content Richness, and Web Interactivity. International Journal of Human-Computer Studies 65.1 (2007): 7184. Web. 16 Jan. 2014.
Scott, John. Trend Report Social Network Analysis. Sociology. Volume 22, Issue 1, (1988): p. 109-127.
Torok, Robyn. Facebook Jihad. A Case Study of Recruitment Discourses and Strategies Targeting a Western Female. International Cyber Resilience Conference. Perth, 2011. 8494.
Wasserman, Stanley and Katherine Faust. Social network analysis. Methods and applications. Structural analysis in the social sciences. Cambridge: Cambridge University Press, 1994.
Weimann, Gabriel. Terror on Facebook, Twitter, and Youtube. Brown Journal of World Affairs 16.2 (2010): 4554.
Waskiewicz, Todd. Exploiting Jihadist Social Networks. 2011 : 132.
Waskiewicz, Todd. Friend of a Friend Influence in Terrorist Social Networks. The 2012 World Congress in Computer Science. 2012.
Black standard - Wikipedia, the free encyclopedia. Wikipedia. January 17th 2014
Facebook. Netvizz. University of Amsterdam. January 17th 2014
NetworkAnalysis < MoM < digitalmethods.net. University of Amsterdam. January 17th 2014Back to top]