group 3.5

Blog Post #1

Use of Statistics in “NSA Files: Decoded”

In the article NSA files decoded by Macaskill and Dance, statistics is used as an important tool to analyze and justify claims.

The first section of the article aims to provide an introduction to Edward Snowden’s influence upon the outlook on data surveillance in our society. A figure provided includes the estimate of the number of friends an average Facebook user has, which is 190 (Macaskill, Dance, 2013). This number aims to provide an insight into how connected users are and into the “three degrees of separation”. This figure is crucial as it lets the reader comprehend that users, are not only connected to their “friends”, but also their secondary friends. Thus, our connectivity to the world is much bigger than anticipated. The authors provide comparative statistics to visualize the provided population statistics. Furthermore, suggesting that NSA is an active examiner of our communications data, the authors have offered a live counter showing the number of terabytes NSA has selected for data review.

In the second section of the article, the author presents the arguments from both sides in order to provide a more completed view for the readers. The NSA claims that they had prevented 54 plots, while the media and congress has drastically reduced this number to four by fact checking. The data here are used in a comparative manner to reveal the lies of NSA, which reinforce its suspiciousness and connotes a sense of distrust between citizens and the NSA. The details of the origins of the data is not given in the text, which reduces its reliability to a certain extent.

In section three, the article provides statistics on the number of fiber-optic cable connections between countries. The article refers to how data collection had boomed as surveillance expanded rapidly due to the fear of terrorism after the 9/11 attacks. The NSA collected data by tapping traffic in the US as much of the world’s communications passes through the US. Besides the US, the NSA also has access to data collected by close intelligence partners such as Britain’s GCHQ. The statistics show that the US is connected to the most countries, being 63, with fiber-optic cables. This data shows that the NSA had an immense data collection system for people not only concerning the US but the whole world. The source of this data is from TeleGeography however the source is not accessible anymore. The data is relatively not prejudicial as it only shows the fiber-optic cable connections between countries. The statistics merely express how much data the NSA would have access to if these connections were tapped.


Section four is relatively short and does not contain any statistics. Contrary to the other parts this one doesn’t serve to talk about data surveillance but rather talks about the future of data protection, namely encryption. It’s interesting that there are no statistics being added here because statistics can be used to predict the future, which would have provided the paper with a variety of types of statistics. Since this section is arguing for the importance of encryption when it comes to data protection a statistic on what percentage of companies or organisations have data encrypted would have added a sense of legitimacy for the future of data protection.

Section five presents the readers with a chart. The data provided in the chart comes from the Judicial Centre which is a relatively reliable origin. It supports the claim about the homogeneity of the Fisa court. The data categorised the judges from the Fisa court by their political background, races and gender. The result shows that the white male Republicans are dominating in the Fisa court. There is also a graph which is illustrated with the names of the judges against the period of time they served. It is a visualization of the data collected which is not an approximate estimation, enabling the readers to understand the power of Fisa court and the potential problems it might exist. It reflects the importance of visualization of data as it highly simplifies the information and conveys the message in a direct manner. However, this data does not directly relate to data surveillance itself as discussed in the rest of the article. It mainly blames the legislation structure of the government and their unfair preference of the choosing the judges.

In conclusion, we have seen the impact Snowden and the PRISM program brought through the article and its statistics. It is obvious that the statistics are a strong tool, that can contextualize our impression, as well as serving as convincing evidence to portray the authenticity of a theory or assumption.

Blog Post #2

“Fbtrex” as a Means of Analysing Filter Bubbles on Facebook

Throughout this week, we conducted a journalistic research on the filter bubble effects on Facebook with the help of the tool “Fbtrex”. Whilst doing this assignment we did not encounter any problems such as our account being deleted by Facebook. The journalistic research we made to some extent lacks authenticity because the persona does not exist in reality which means that it did not simulate a real Facebook account. The friends added were not based on any realistic relations either. Thus all  activities on Facebook cannot be considered valid and might be problematic especially when the friends added are ‘real’ while the persona is ‘fake’. Therefore the ethics behind creating a ‘fake’ account is questionable however as long as the account is not used to deceive anyone, the fake account cannot be considered harmful. Especially as the only intention behind the creation of the ‘fake’ account was to investigate the filter bubble mechanism of Facebook.

“Fbtrex” is a unique tool which allows us to observe Facebook from different points of view and understand the mechanisms behind the algorithms (why we see what we see, who sees what, etc.). The data that are collected by the tool are public posts which appear on our newsfeeds. If a certain post has been recorded by the tool, it says “This post has been recorded” above the post itself. Furthermore, the tool only collects public information on our timelines and does not invade other personal data such as our messages. Thus, “Fbtrex” can act as  a helpful tool for understanding how filter bubbles work and how news are distributed.

In regards to the ethical implications of the tool fbtrex, we thought that certain individuals could easily find this tool ethically invalid and feel as though their privacy is invaded and monitored. Though it is made evident that the tool is used for the mere purpose of observing the social network from different viewpoints and comprehend Facebook’s algorithmic mechanisms, users can feel as though their actions are being watched whilst using this tool.

Our fake persona is called Clayton Reed, a 27 years old British male. He is a graduate from London South Bank University and as a group we decided that politically he stands for Brexit as a right-wing political supporter. As for his hobbies, we decided that he enjoys television shows such as How I Met Your Mother and Peaky Blinders. He is passionate about playing frisbee and his favorite type of music genre is R&B and Hip Hop. To legitimize and analyze this ‘fake account’ for six days, we liked and shared pages on his timeline. Our objective was to see how pro Brexit contents he liked affects the filter bubble of Facebook. An example is on day 5 Clayton liked a journal article recommended by Facebook, and the article was posted by Daily Express, titled ‘EU is ‘no longer major power’ says from Barclays boss’. The notifications received in this account are mainly the questions and surveys sent by the Brexit communities. Through this you can see we were actively engaging in pages directly in line with the beliefs we created. It also reveals the filter bubble effect of Facebook which only enables its users to see the contents they are interested in and blocking out all the other counter perspectives. The public discussion is limited and Facebook users are trapped in their own sphere without receiving any information that may expand their view.

 Group 5: Emilie, Quiting, Ilaya, Irem, Dylan


By: Irem, Qiuting, Ilaya, Emilie, Dylan

Posted: 20 Mar 2019

Amsterdam, the Netherlands – The objective of our research investigation was to investigate the influence of filter bubbles on pro-Brexit users on Facebook. Thus, we created a fake persona on Facebook by the name “Clayton Reed” and made sure that he was following a variety of pages, adding various people and liking posts. With the help of the Fbtrex tool, we were able to view an Excel spreadsheet which allowed us to see which pages appeared most frequently on Clayton’s newsfeed, what kind of posts they were, and which categories they fell into.

We have created two individual pie charts in order to illustrate our findings from the research investigation. The first pie chart, labelled Figure 1 below illustrates the distribution of the formats of posts received. It is evident to see that the majority was labelled as “posts”, landing at a percentage of 54.43%. 29.24% were photos, 10.07% were videos, 4.97% were posts from groups and 1.38% were events. The reason why the majority of the posts seen on the newsfeed were labelled as “posts”, could be because Clayton was an active followers of news and magazine outlets. Moreover, Clayton did not have many friends, which means that he was not exposed to a variety of images and videos.



We decided to focus on the posts specifically as it contained the authors that appeared most frequently in Clayton’s feed with 54.43%. We decided on the top five most frequently appearing authors as there were more than a hundred different authors just within posts. What we found that WIRED Magazine appeared most frequently covering 12.47% of the feed. Quartz appeared second most often at 6.87% and Daily Express third at 4.58%. The Source Magazine and La Cite de la Voix – Vezelay both appeared 3.31%. Since WIRED Magazine appeared the most often we can say that most of Clayton’s feed appeared slightly left-wing even though his feed actually remained quite neutral.



Within all the data of the authors we picked the 5 most popular ones to analyse and further research. The majority of the authors were politically neutral, or more specifically had points of view from both sides of the political spectrum. We disclosed that the reason for this was that lots of magazines aim to be neutral and provide different points of view. One of the magazines Daily Express was right wing and the magazine WIRED is left wing. Based on this data we can disclosure that our persona had a relatively neutral feed.

In regards to the ethical implications of the tool “Fbtrex”, we thought that certain individuals could easily find this tool ethically invalid and feel as though their privacy is invaded and monitored. Though it is made evident that the tool is used for the mere purpose of observing the social network from different viewpoints and comprehend Facebook’s algorithmic mechanisms, users could easily feel as though their actions are being watched whilst using this tool. Furthermore, there are also ethical complications of creating a fake persona on Facebook. Firstly, we have disregarded Facebook’s regulations and the fact that it is not allowed to pretend to be someone else. Moreover, we have ultimately used someone else’s photography without having their permission, nevermind the fact that the photo is from Shutter Stock.

Moreover, there were several limitations to this particular research investigation. For instance, we were faced with several posts which were labelled as “Null”. These posts did not fall into any category thus were eliminated from the analysis. This meant that we had data points which were not of use to our research and thus had to be ruled out. Furthermore, since we only had one week for the research, the data gathered might not be sufficient enough to make a valid analysis about the subject.

In conclusion, we believe that this research project’s had a very limited time frame. It was too short of an experiment, thus our fake persona did not have a search history (only about 6 days, and not very active even then), no friends (not a lot of people accepted us since they did not know who Clayton was) and he had no experience. However, from the experiment that we have conducted, the posts visible on the newsfeed seem to be neutral. The posts were not directed merely for pro-Brexit users, and the posts that we could see on the newsfeed had taken a neutral stance on the Brexit issue. Moreover, we tried to join several inclusive pro-Brexit groups, however the ones that we tried to join did not let us in. In addition, it was difficult to find groups that were labelled as pro-Brexit since this is a highly specific audience to find on Facebook.

Data Journalism Final Report Link to PDF:

Data Journalism Final Report