Over a period of analysis from December 15th, 2018 to March 15th, 2019, Alto’s team of data scientists conducted an analysis of the public, digital sociopolitical conversation related to the EU elections in Germany. Public data analyzed includes over 9,650,000 million results from 755,787 authors generating 552,617 conversations and sharing 783,954 pieces of content across multiple digital sources such as news sites, Twitter, Facebook, YouTube, Instagram, blogs, forums, and other digital communities. The goals of the analysis were:

  1. To identify key narratives and communities of opinion, the users that make up these communities, and the issues addressed by these users as a means to identify any abnormal behavior that might signal or point to disinformation or disinformation strategies.
  2. To identify users who demonstrated abnormally high activity during the period of analysis and the impact of their activity across the broader network of users.

Through this analysis, Alto’s analytics team derived the following insights:

  1. Our analysts continue to observe a high level of polarization in the public digital conversation. Content analyses yield several examples of disinformation aimed at influencing the public debate.
  2. A group of 168 (0.07%) users with abnormally high activity generate at least 25 and up to 197 posts per day, resulting in a total of 650,971 (10.77%) posts over the period analyzed. This finding confirms the previously identified patterns of a small number of users (around 0.1%) creating over 10% of the observed public results.
  3. These users are primarily composed of authors demonstrating an affinity of interaction for the AfD Supporters community and thus identified within the AfD Supporters community. 76% of abnormal high activity users who produced 78% of comments are identified as part of this community.
  4. The analysis has identified the top 6 salient narratives in the public debate which range from generic discussion about political parties and their agendas to EU legislation on copyright laws (Article 13), Climate Change and environmental issues, and wider discussion on international affairs including the debate on Iranian refugees in Turkey.
  5. Several sites that generate highly segmented content, focus heavily on issues such as immigration and Islam, and bear anti-EU discourse are shared with high frequency by the AfD Supporters community. Several of these sites are among the most relevant domains over the period of analysis.
  6. Analysis of traditional-mainstream, emerging, and foreign media in the debate and in the creation of the key narratives provide confirmation of the increasing relevance of both emerging digital sites and foreign media sites. Beyond the key role of international global services and platforms such as Twitter, Facebook, Instagram, WordPress, Google, or YouTube, the analysis shows how other media sites play a key role. Russia Today (RT or rt.com), Change.org, Anonymousnews.ru, Sputnik (sputniknews.com), Wikipedia.org or Vice.com are within the top 100 most influential sites. Gab.ai, known for attracting far-right and alt-right users who have been banned from other social networks, ranks in position #101. Russia Today (RT or rt.com) appears among as one of the most shared domains in the debate at 29th, increasing in relevance from Alto’s previous analysis of the German public digital sphere conversation from December 2018 to January 2019 when it ranked at position 46th.

Narratives Analysis

Alto’s software analyzes the high-frequency terms, keywords, and hashtags used in public posts by users or by journalists in articles and visually connect them when they are mentioned simultaneously, or in the body of the same text, by the same users. This method enables unassisted and unbiased identification of key emerging narratives and their complexity and richness through assessing the density of unique terms used within each narrative. The graph below shows the top 208 keywords & hashtags representing 48.7% of the total results analyzed. Alto’s analysts have clustered the different narratives using AI algorithms that identified the following main themes in the public digital sociopolitical debate.

The following six key narratives emerged from the analysis:

  1. AFD related (47.1% of high frequency terms): Topics on policy and social issues with the AfD party are at the center of the conversation. Both supporting and detracting narratives are visible with prominent hashtags such as #nazisraus, #islam, #afd, and #noafd, among others.
  2. Other political parties, national issues, and European elections(23.1% of high frequency terms): This conversation is focused on national politics and matters pertaining to the European elections. Frequently used hashtags include #spd, #cdu, #csu, #Europa, #deutschland, #fdp, and #grüne, among others.
  3. Article 13 (12.5% of high frequency terms): Includes key terms pertaining to the recently introduced EU legislation on copyright laws. Salient hashtags include #artikel13, #uploadfilter, and #saveyourinternet, among others.
  4. Climate change and environmental issues (5.3% of high frequency terms) Conversation focused on climate change and environmental matters. Key hashtags include #bundesregierung, #fridaysforfuture, and #klimaschutz, among others
  5. Iranian refugees in Turkey (5.3% of high frequency terms): Merkel and Germany make up much of the conversation centered on Iranian refugees in Turkey. Notable hashtags include #merkel, #Germany, #iranianrefugeesinturkey, #humanrights, and #un, among others.
  6. International issues (3.9% of high frequency terms): The Yellow Vests in France and the Venezuelan political situation are key topics of discussion. Frequently used terms include #macron, #giletsjaunes, and #venezuela, among others.

Communities Detection Analysis

Alto data scientists used Alto Analyzer, Alto’s proprietary cloud-based analytics platform, to build and identify the largest communities of users interacting in connection with the German public debate. This community was identified in Twitter, and their interactions were filtered for retweets only, in order to perform a topological analysis of the propagation of their messages. This resulted in a unimodal network of authors (giant component, edges being re-tweets). To determine the different communities and visually represent the resulting network, Alto’s data scientists applied clustering algorithms, like the Louvain Method for community detection, for example. The visualization below represents the most relevant 232,304 users and their 4,120,602 interactions.

Network topology is characterized by network structures with nodes (profiles) and edges (lines) which connect them. Alto’s algorithm clusters nodes and edges in order to calculate proximities and communities of users.

The following five key communities emerged from the community detection analysis:

  1. Left community (36.9% of users, 34.2% of interactions): Criticisms of the FDP, CDU, and discussion around racism against refugees and immigrants. Conversation in this community includes support for the #Nazisraus movement. Among the most influential profiles are Die Grünen and Die Linke party profiles.
  2. Piraten Partei (25.5% of users, 8.9% of interactions): Criticisms of the EU’s copyright legislation, Article 13, and criticisms of the CDU.
  3. CDU and SPD (18.2% of users, 8.3% of interactions): General comments on national issues and claims for democracy in Venezuela. Includes criticisms of the “hypocrisy” of Die Grünen party.
  4. AfD Supporters (10.7% of users, 46.3% of interactions): Spread of the attack on Frank Magnitz (AfD politician), criticisms of immigration and criticisms of other parties such as Die Grünen and the CDU.
  5. Solidarity with Catalan Independence Movement (3.7% of users, 2.3% of interactions): Bavarian nationalism profiles expressing solidarity between Bavarian nationalism and Catalan independentists are the most influential authors in this community.

Politicians Among the Top Communities

In an analysis of influential nodes in the communities mapped, Alto measured influence mathematically based on the structural connections of the authors in the digital debate and their ability to propagate messages to larger audiences in the network (centrality measurement). The results of this analysis demonstrate that out of 72 users identified as verified politicians or political parties (0.03%), 21 are from the SPD, 16 are from the CDU, and 15 are from Die Grünen. Regarding the communities in which these politicians and political parties were identified, as can be seen below, 45 politicians were identified as authors with affinity to the CDU and SPD community, more than any other community. 21 verified politicians or political parties were identified as demonstrating affinity to the Left community, while 6 were identified within the AfD Supporters community.

Abnormality Detection – High Activity Users

Alto’s analysts identified a group of 168 (0.07%) users demonstrating abnormally high activity. These users produced a total of 650,971 (10.77%) posts over the period of analysis. These users are found to be concentrated in the AfD Supporters community — 76% of abnormal high activity users who produced 78% of comments are identified as part of this community.

High Activity Users Generate Significant Activity

The abnormal high activity users manage to attract and engage a significant volume of other users. This extended network of users interacting with the abnormal high activity users through retweets, replies, and mentions produce 928,421 total interactions over the period analyzed, demonstrating an effective pattern of driving the conversation beyond only retweets and amplifying engagement. The graphic demonstrated below represents unique retweets, mentions, and replies among users interacting with the abnormal high activity users in the analysis. Users interacting via multiple forms of engagement with the abnormal high activity users, or any combination of retweets, mentions, or replies, make up the remainder of the interactions, or those not counted as unique interactions.

Abnormality Detection – Account Creation Dates

36.3% of the profiles of the 168 abnormal activity users were created since 2017. This is consistent with a trend identified other Alto analyses. Of the high activity accounts identified in the public digital sociopolitical debate, a significant proportion have been created within the past one to three years.

High Activity Users – Most Shared Content

Role of Traditional, Emerging, and Foreign Media in the Debate

To determine which media sites are most influential, Alto’s algorithms take into account the number of users and the volume and frequency of links from media sites being shared across social media sources to calculate how influential a site is within the context of the analysis. Alto’s software ranks content influence similar to Google’s algorithms for web page ranking – the content from more relevant sites are likely to receive greater attention from more users over a sustained period of time.

Alto’s analysis found that traditional German media and social media platforms ranked high in influence, as did a growing network of emerging digital media with highly segmented and specific editorial strategies. Please see the table below for the top 150 most influential domains and the contribution of each community to the propagation of each domain’s content within the network. In an ideal scenario, content from mainstream media would tend to be distributed in a balanced way across different communities, while highly segmented content would demonstrate a tendency of higher consumption by their key communities of interest. However, given the high polarization of the debate and the weight of high activity users, the table shows how almost no mainstream media achieves that ideal scenario.

Some of the key findings are as follows:

– As identified in Alto’s initial analysis of the German public digital sphere conversation, the AfD Supporters community has a high level of affinity to several sites which rank among the top most relevant domains and are shared only marginally by other communities identified in the analysis, likely due to the high level of segmentation with a specific angle on issues such as immigration and Islam which characterizes their editorial content. Some examples of these sites, as can be viewed in the tables below, are journalistenwatch.com (99%, Affinity to AfD community), bild.de (89%, Affinity to AfD community), jungefreiheit.de (98%, Affinity to AfD community), tichyseinblick.de (93%, Affinity to AfD community), epochtimes.de (97%, Affinity to AfD community), philosophia-perennis.com (99%, Affinity to AfD community), pi-news.net (99%, Affinity to AfD community), ffd365.de (100%, Affinity to AfD community), afbundestag.de (99%, Affinity to AfD community), anonymousnews.ru (98%, Affinity to AfD community), all appearing among the top 50 most relevant domains in the study.

– 18 of the top 50 domains are shared at a rate of over 80% by the AfD supporters community, demonstrating a tendency to share specific domains with high intensity as a means to raise awareness on their key narratives of interest.

– Facebook (82%, Affinity to AfD community) is frequently shared by identities within the AfD community, exemplifying the importance of content from Facebook within this community.

– Russia Today (RT or rt.com) and Sputnik (sputniknews.com) appear among the most shared domains in the debate at 29th and 55th, increasing in relevance from Alto’s previous analysis of the German public digital sphere conversation from December 2018 to January 2019 when they ranked at 46th and 69th, respectively. In the current analysis extended from December 2018 to March 2019, RT is shared with a frequency of 74% and Sputnik is shared with a frequency of 82%. Although shared at a lower rate in the AfD Supporters community than in our previous analysis, these sites increase in overall relevance over the period analyzed. This signals the continued presence of Russian media in the German socio-political debate.

Over a period of analysis from December 2018 to January 2019, Alto’s data science team conducted a previous public digital sphere analysis of the German sociopolitical conversation. Read the analysis here: Germany: Digital Public Debate Ahead of EU Parliamentary Elections

Interested in our work? Please contact us at info@alto-analytics.com. To learn more about Alto Analytics, subscribe to our newsletter below.