Memory conflicts on social media: a Twitter data analysis
The second report (2019), by Celeste Muñoz Martínez
Over the course of this tracking analysis of the signification and resignification of Europe’s past on social media, technical and content reports have been prepared on Twitter user activity surrounding commemorations and other key dates in order to obtain a broad perspective, collect a significant number of data points and gain the ability to make meaningful comparisons. As a result, the present report offers both quantitative and qualitative analyses.
At the wider European level, three dates have been selected for analysis: Holocaust Remembrance Day (27 January), Europe Day (9 May) and Remembrance Day for Victims of Totalitarianism (23 August). These dates correspond to annual commemorative events. Given that they are Europe-wide in nature, their evolution has been easier to monitor. However, one-off events (i.e., centenaries, anniversaries, etc.) have also been incorporated because they hold special interest. In this vein, the sample includes the centenary of the end of the First World War (11 November, in 2018), the 75-year anniversary of the D-Day landings at Normandy (6 June) and the 30-year anniversary of the fall of the Berlin Wall (9 November). In total, the Europe-wide portion of the analysis has addressed six dates from late 2018 through the entirety of 2019.
At a specifically Spanish level, the sample additionally contains a date that is not officially recognised as a festivity, but that is central to the construction of memories regarding the Spanish Civil War and the Franco dictatorship: the commemoration of the Second Spanish Republic (14 April), which was cut short by the outbreak of war and subsequent Francoist regime. In this way, the analysis has been able to look at official vs. informal festivities, give some consideration to state memory and related policies, and reflect on movements toward acceptance, rejection or vindication on social media. In addition, two other Spain-related dates have been included: the day of the liberation of the Mauthausen concentration camp (5 May) and the unfolding controversy surrounding the police headquarters in Via Laietana in Barcelona after a plaque was put up in remembrance of the repeated torture that was carried out there.
Table I. Sample overview
|Police headquarters in Via Laietana, Barcelona
Commemoration of the Second Spanish Republic
|27 March to 01 April
Day of the liberation of Mauthausen concentration camp
75-year anniversary of the D-Day landings at Normandy
|Remembrance Day for Victims of Totalitarianism
30th anniversary of the fall of the Berlin Wall
|Centenary of the end of the First World War||9-12 November|
The first step was to identify the trends associated with each commemoration using trending topics (henceforth TT) and keywords (without hashtag). The aim was to follow the digital trace of a particular phenomenon, seeking to include the greatest number of multi-situated trends, which can be expressed in many different forms within each community. For the present analysis, the portal Tendrinalia was used. This has made it possible to collect viral trends (by country, city or date), primarily in France, Britain, Spain and Germany, but also to a lesser extent in Italy and Portugal. The digital expressions of these countries have been particularly interesting in relation to their linguistic features because the TT and keywords in the lingua francas of these states spread subsequently to surrounding states as a result of their greater public presence (see Figure 1 below). However, despite the fact that trends in Bulgarian, Polish, Slovenian, Hungarian, Estonian, Russian and other languages have been taken into account for the two samplings on 23 August and the fall of the Berlin Wall, it nonetheless remains a future goal to obtain a better representation of Twitter use in Eastern Europe.
At this point, it must be stressed that the language factor does not indicate precise geographical trends. Rather, it shows the amount of data at a macro level that corresponds to a language community, which may be limited to one country or involve many countries. Consequently, the analysis must take into account that tweets in English reflect a much more geographically undefined phenomenon. Similarly, it is necessary to consider a large number of Spanish speakers located in Latin America and a large number of French speakers in Africa and Canada. However, to limit the digital trace to Europe as strictly as possible, the analysis has focused on hours of use. In other words, any online activity while it is night in Europe reflects heightened activity in the United States or Latin America, an aspect that is also represented in the display of data (see Graph 1 below).
Figure 1. Sample of the volume of tweets by language during the commemoration of the fall of the Berlin Wall (2019). The relationships among languages show that language is a barrier to trending topics or debates. English, followed either by German or Spanish, has the most interactions with other language communities. It must also be noted, however, that the demographic superiority of certain communities makes it hard to represent and make visible any languages that have fewer speakers.
Graph 1: Sample of the temporal representation of the spread of tweets by day, time and language during the commemoration of the 75-year anniversary of the D-Day landings at Normandy (2019). As the graph shows, the predominant languages were English and French, whose use was comparable in the volume of data throughout the first part of the day. During the second half of the day, however, activity in French fell while there was a sharp upturn of activity in English corresponding to users in the United States. This phenomenon recurs on 9 and 10 June, when the time bands with the greatest volume of data occur during U.S. hours of use.
At a technical level, two methods have been used to collect the data appearing in the figures and graphs: the first is web scraping, which extracts a file with metadata on users and their content and makes possible a broad chronological scope that may even be retroactive, while the second is T-hoarder, which uses an API (search or streaming) to generate displays of short- and medium-term phenomena in real time using search filters. By combining the various possibilities offered by the two methods, it has been possible to produce approximations both in detail and in general. However, the scientific criteria for data collection have been modified or expanded as a function of the conclusions drawn from the initial samples, which were markedly quantitative in nature. Subsequently, in the second half of the project, improvements at the technical level and in content analysis were employed to undertake a more thorough sampling of the most relevant tweets (between 10 and 20 by TT), with the aim of not basing the outputs solely on macro data that are hard to display on discursive axes.
Lastly, it was necessary to define the technological and personal profile of users based on statistical data, understanding that the opinions and phenomena on Twitter fit within a clear-cut sociological segment and a particular function, in order to produce a better comparative analysis that is also well-situated. In addition to these exogenous analyses, which are nonetheless central methodologically, it would also be advisable to look in greater detail at European and Spanish legislation on online hate crimes, which can have a strong inhibiting effect on the expression of some positions on Twitter.
The sociological and technological profile of users
In 2018, 21.8% of Twitter users accessed online content from Europe. To be exact, this figure breaks down to 14.2% from Western Europe and 7.6% from Central and Eastern Europe. In population terms, the figure represents a total of 56.5 million Europeans, given that in 2018 there were 259 million users on Twitter (which is more firmly established in the United States). Twitter as a platform has not stopped growing since it was founded in 2006: for example, the number of profiles has risen over nine years from 20 million profiles to 262 million profiles in 2019 and its future prospects point to continued growth. For these reasons and because of its ability to generate content and opinions, Twitter has become a target in recent years for phenomena like election campaigns, politicians and fake news. As a result, it is necessary to gain some idea of the prototypical public that interacts on the platform in order to situate the analysed trends within a specific stratum of the population.
First, Twitter users are primarily men (66%), while only 34% are women. This makes it one of the most masculinised social networks. In addition, beyond the wide gender difference, Twitter profiles correspond generally to young people (53.6% of users are between 18 and 34 years old) or middle-aged people (21.7% are between 35 and 49 years old). In addition, users access the platform largely with their mobile telephones, live in urban areas, have a high technological profile, and have attended university. It is also important to bear in mind how users make use of the platform. While Twitter has considerably fewer users than Facebook or Instagram, its main function is 56% informational, unlike the other social networks, which are largely employed to share photos and personal content. In this respect, Twitter is a much more important arena of politicisation.
At the European level, however, the popularity of Twitter varies by country. For instance, the United Kingdom (17.1 million users) and Spain (4.1 million users) are the top two European countries by volume of activity on the platform. Both countries figure in the top 10 worldwide by the number of users and together account for a third of all profiles in Europe. By contrast, Germany is the country in Europe, and practically in the entire world, that has the lowest level activity on social media. Accordingly, one goal for the future will be to conduct a more specific and focused study by country in order to gain a better understanding of online phenomena and gaps.
Analysis by event
Centenary of the end of the First World War
To understand the data on this commemoration better, it is first necessary to look at the controversy sparked in France in the preceding days, which largely drove the inclusion of the date retroactively (using web scraping) even though the event precedes 2019. On 11 November 1918, representatives of the Triple Entente on one side and the German Empire on the other side signed the armistice, known as the Armistice of Compiègne, that ended the First World War. During advance preparations for the centenary, the President of France, Emmanuel Macron, raised the possibility of paying tribute to Marshal Pétain, who had been hailed as a national hero for his role in the Battle of Verdun in the First World War, only to become, during the Second World War, a key collaborator with Nazi Germany as leader of Vichy France. In addition to this controversy, the historical significance of the armistice itself, together with the historical and socio-political consequences of the war, helps to understand why the centenary became such a viral phenomenon with as many as 148,489 tweets (which reached 547,542 through retweets).
The hashtags tracked on the day for analysis were multilingual in order to increase representativeness. Prominent examples included #WWI, #Armistice2018, #ww1centenary, #ErsterWeltkrieg, #Centenaire1418 and #PrimeraGuerraMundial. By far, the most frequently used hashtag was #ArmisticeDay100 (at 12.98%). As is common, however, the largest amount of content was generated without a hashtag (21.31%). As a result, the searches also covered key words in all languages (see Graph 2 below). The generation of content was most heavily concentrated on 11 November, when there was a high level of activity during the hours of the night in Europe, coinciding with high activity in the United States (where the most common hashtag was #VeteransDay).
By examining all of these patterns, the language communities have been identified. English shows the greatest influence (66.98%), followed by French (17.22%) and Spanish (8.28%). Also notable is the very residual presence of German (0.45%).
Graph 2. Twitter activity by TT during the commemoration ceremonies and preceding days.
Within language communities, it has also been possible to identify geographical subgroups, which permit a situated analysis of the trending topics and their agents. More specifically, the largest share of interactions among groups involved content generated in the United States, the United Kingdom, and France, whereas the remaining groups were somewhat affected by the language barrier. However, the connections must also be accounted for historically, given that the three countries in question fought on the same side in the war and they reproduce this dynamic in their online patterns (see Figure 2 below).
Figure 2. Distribution of connections by community.
The data on this commemoration correspond to the project’s initial reports on Twitter, which were then used to refine the search criteria. At the level of content and discourse, therefore, it has not been possible to include a data readout, given the insufficient data points needed to plot the reach and main agents.
International Holocaust Remembrance Day
In 2005, the General Assembly of the United Nations and the European Parliament established 27 January, the anniversary of the date on which the Soviet Army liberated Auschwitz in 1945, as an international day of commemoration in memory of the victims of the Holocaust. In relation to 27 January 2019, the data collection effort has gathered a large amount of relevant data to facilitate more situated analyses. The Holocaust remembrance day is a commemoration of special importance and interest because of the political interpretations, sensibilities and instrumentalisations to which it has been subject throughout history and today. In this respect, the aim of the study has been to situate the main arguments and lines of discourse. To begin with, the sample and its reach have been much broader than in the previous example, achieving a total of 26,954 tweets (which rose to 123,776 through retweets). At a methodological level, web scraping has been combined with API monitoring, especially between the days of 25 and 30 January, resulting in improved sampling.
One of the key conclusions is that many communities, both small and dispersed, appear clustered into affinity groups, showing that it was a commemoration in which many less connected sensibilities nevertheless interacted with one another (see Figure 3 below). The predominant language of use was English (70.02%), followed far behind by Spanish (20.12%), French (4.2%), German (2.9%) and Catalan (0.84%). As indicated earlier, the language was a barrier to interactions among communities, to which ideological affinity could also be added: as Figure 3 below shows, English-speaking and Spanish-speaking users had few interactions between one another, with Catalan being the language that had the most relationships with Spanish and German being the one with the most relationships with English, among those shown in the figure.
Figure 3. The table on the left indicates affinity communities and their respective percentages. The data show the disaggregation and unrepresentativeness within the affinity communities, which indicate that the conversations were dispersed and located in small groups. The diagram on the right indicates the connections and interactions between language groups.
At the content level, the largest share of the tweets was again generated without a hashtag (36.89%) and they have been collected using keywords, specifically Holocaust, Holocaust and Holocausto. At the same time, the hashtags that brought together the most content were #WeRemember (16.01%), #HolocaustMemorialDay (8.27%), #HolocaustRemembranceDay (6.82%), #Holocaust (3.62%), #HDM2019 (2.99%) and #Holocausto (2.05%).
The largest amount of content was generated initially by media outlets, international or political organisations, and public figures. These include the British organisation Holocaust Memorial Day Trust (@HDM_UK), the U.S. organisation for the civil rights of LGBTQ groups (@HRC), the official Twitter channel of the State of Israel (@Israel), the British artist Stuart Humphryes (@StuartHumphryes), the President of the United States Donald Trump (@DonaldTrump), the Speaker of the U.S. House of Representatives Nancy Pelosi (@SpeakerPelosi), UNESCO (@unesco), the regional ambassador for the Holocaust Educational Trust Barnabas Balint (@BarnabasBalint), the artist group Inaccessible Art Fair (@inartfairNYK), the social psychologist Petra Boynton (@DrPetra), the Portuguese artist Armanda de Andrade (@ArmandaAndrade), the Polish activist Kasjopea – Anita Schelde (@AnitaSchelde) and the Israeli organisation Stand with us (@standwithus). However, leaving aside the content’s origin and the fact that most of the content commemorated the day with messages of peace, memory, photos or personal stories from relatives of the victims of the Shoah (for instance, users such as the above-mentioned @inartfairNYK or the writer Nadine van der Velde (@nadinevdVelde) “#WeRemember It is my duty to remember. This is the last picture of my family in Apeldoorn, Netherlands before they were terrorized by Nazis, went into hiding, murdered in concentration camps or enslaved in labor camps My father is the boy in the center. #HolocaustMemorialDay”), the interest here is to highlight the presentist use of the events by less prominent profiles (who also have less reach). The aim is to analyse at a popular level, within the sociological segments in question, how the past is incorporated into the present, what stories are told and to what end. For example, many users have used the commemoration to attack Donald Trump and to draw parallels with his policy on borders and immigration (e.g., the satirical fictional character Mrs. Betty Bowers (@BettyBowers) tweeted “#HolocaustMemorialDay is an appropriate occasion to reflect on this nation’s history of “America First” and hostility to refugees fleeing oppression. And to wonder how we will look to future generations” [930 RT]). The commemoration was also used to criticise Brexit (the physician Edzard Ernst (@EdzardErnst) tweeted “#HolocaustMemorialDay – a very good day to contemplate what #Brexit is doing to our nation, I think”) or to condemn the rise of the far-right (@climate_earth: “I believe the World is seeing the rise of fascism again with #Trump in the White House, #Brexit and the growth of far-right parties from Brazil to Europe. Unless we heed the warnings on #HolocaustMemorialDay it could conceivably happen again. #FBIR #FBPE #facciamorete”). On more than a few occasions, the responses also included tweets under the hashtag #ArmenianGenocide, which had happened as well in the analysis of the centenary of the end of the First World War, in order to raise awareness of this poorly handled aspect of the past, or under other hashtags that made reference to the Romani people or LGTB groups, who were additional targets of the repressive policies of the Third Reich. In other words, Holocaust Remembrance Day witnessed a confluence of many political and social memories, analytic perspectives, messages (including denialists) and uses that took the shape of lessons, critiques or debunking. The commemoration offers an example of how the significations of the past can take on different forms in the present and serve to draw connections or pursue revisions that will be used later by data analytics firms to craft messages individually tailored by institutions or politicians for use in election campaigns or for use in the commemorations themselves. In this respect, in-depth analyses of Twitter can help to identify reach and prevent or counteract populist trends that are now gaining ground on social media.
Europe Day, which is celebrated on 9 May in memory of the 1950 Schuman Declaration, is a date that seeks to build a bond of shared identity across the entire European continent. It also appears in trending topics on social media, where it is used by the European Commission and its online circles, as well as political parties. However, it lacks deep social engagement. Specifically, between 6 and 13 May 2019, the data collection effort has used a combination of web scraping and API monitoring to find 24,222 tweets (with a reach of as many as 114,447 through retweets). For this sampling, however, it has not been possible to include content without hashtags (which is normally greater in number), because the words “Europe Day” or anything similar may give false positives linked to football. Accordingly, the search focused on the day’s most influential hashtags, such as #EuropeDay (27.75%), #DiaDeEuropa (9.31%), #DíaDeEuropa (8.18%), #Europatag (4.08%), #EuropeDay2019 (3.33%) and #JourneeDelEurope (1.9%).
Indeed, the two diagrams in Figure 4 below demonstrate perfectly the formation of affinity communities strongly linked to institutions or political parties within each language group. The predominant language group is once again English (36.21%), followed by Spanish (31.85%), French (13.61%), German (6.7%), Italian (3.51%) and Catalan (2.02%). It has also been possible to identify subgroups linked to political parties or institutions that promote the social phenomenon within each language community. As can clearly be seen in the Spanish-speaking community, for example, Unidas Podemos, PSOE and Ciudadanos were among the leading drivers, whereas the European Commission and related circles gained prominence on social media using English (see Figure 4 below).
Figure 4. Spread by language community (left) and by affinity group (right).
In the content analysis, four of the ten tweets with the greatest impact were put out by the European Commission (up to 3772RT). The remainder came from public figures, such as the climate activist Greta Thunberg (@GretaThunberg), who took the opportunity to recall the role that Europe must play to address the issue of climate change (2141RT), and Polly Polak, one of the voices that were most critical of Brexit (@PolakPolly: “Today is #EuropeDay, a day to celebrate peace and unity in Europe. The million+ EU citizens in London are our friends, neighbours, and colleagues”) (643RT). The Spanish tweets with the greatest impact were put out by politicians in the three parties mentioned above, including tweets from the leader of Unidas Podemos, Pablo Iglesias (@Pablo_Iglesias), to Vicente García Riestra, the last living Spanish survivor of Buchenwald (1570RT), and from the leader of Ciudadanos, Inés Arrimadas, on the Catalan political conflict (@InesArrimadas: “Hoy #DíaDeEuropa me encuentro una nueva pintada en mi casa con lazos amarillos señalando donde vivo. La UE se fundó para luchar contra el nacionalismo. Algunos no han aprendido nada”) (1505RT).
Another important tweet came from the NGO Open Arms (@openarms_fund) to recall the refugee crisis and criticise the role of the European Union (511RT). In French, the largest share of tweets were generated and spread by institutions and media outlets. However, the French tweet with the greatest impact over the days in question was posted by Porcher Thomas, a prominent professor of economics, who took the opportunity to criticise European austerity policies in relation to Greece (@PorcherThomas: “En cette #JournéeDeLEurope, rappelons qu’au nom de l'”UE”, la Grèce s’est vue imposer une cure d’austérité au coût social sans précédent: – 30% pr le revenu des ménages, +42% pr la mortalité infantile, +44% pr les suicides,etc. Depuis 2011, l’UE a brisé des millions de vie.”) (1092RT). In German and Italian, on the other hand, the trending topics relating to the date had less impact and were merely institutional in nature, making reference to the founding of Europe and the need to strengthen it and even recalling that the European elections were to take place on 26 May and that it was important to vote. In this respect, a tweet came from Heiko Mass, a member of the Social Democratic Party of Germany (@HeikoMass: “Heute ist #Europatag! Das ist ein guter Zeitpunkt, den Mund aufzumachen: gegen die ewigen Angstmacher. Und auch, um mit der Familie, mit Freunden und Bekannten darüber zu sprechen, warum es so wichtig ist, am 26”) (109RT). Ultimately, Europe Day gave rise to a variety of political uses that varied by context, but originated primarily among political parties and public figures, as is the case with all the commemorations in the sample in terms of their present use and their ongoing collective and socio-political resignifications. However, it must also be said that the impetus behind the festivity is merely institutional in nature, marked at present by very little knowledge or social interest on Twitter.
75th anniversary of D-Day and the Battle of Normandy
The commemoration of the D-Day landings at Normandy is a key date because the year 2019 witnessed the seventy-fifth anniversary of an event celebrated on both sides of the Atlantic, which marked a turning point in the Second World War and led to Allied victory in France. The event’s institutional celebration took place on 6 June, but its impact can be seen on Twitter between 3 and 10 June. As a result, the process has been monitored in its entirety. Specifically, between 3 and 10 June 2019, the data collection effort has located 36,272 tweets (with a reach of as many as 121,148 through retweets), obtained through a combination of web scraping and API monitoring. The search has focused on the hashtags with the greatest influence on the day, nearly all of them appearing in English or French, including #DDay75years (13.82%), #DDay75 (9.36%), #DDay75thAnniversary (8%), #Dday (4.83%), #6juin1944 (3.06%) and #Normandie (2.42%). In keeping with the trend, however, the largest share of tweets did not have a hashtag (25.62%), so the sampling has also covered key words (Normandie, Normandy, Normandía) (see Graph 3 below).
For this commemoration, the two language groups with the highest amount of activity have been English (47.2%) and French (34.99%), followed far behind by Spanish (8.11%), Dutch (2.36%) and German (2.25%). In addition, the geographical distribution shows that the largest number of tweets were written in France, the United Kingdom, and the United States and that the language barrier was not an obstacle to making connections between regions during the commemoration, especially between France and the United Kingdom (see Figure 5 below).
Figure 5: Distribution by language (left) and by region (right). The two diagrams show clearly that the boundary between French and English witnessed many interactions, particularly between the U.K. and France, which helped to make the language barrier more permeable.
Graph 3. The activity of hashtags and TT on the timeline, showing an upturn in activity in the United States when it was night in Europe.
While it has not been possible in this case to define subgroups or ideological and institutional communities, a variety of trends can be seen in the content of tweets with the greatest impact. The most widespread tweet was posted by the U.S. sports reporter Marty Smith, in which he recalled the men who fought in the battle (@MartySmithESPN: “75 years ago, Operation Neptune dropped 160,000 brave young men on the French Coastline. Omaha Beach. Normandy, France. When that gate dropped from that boat, those men jumped straight into hell. We are FREE. We can go. Do. Be. Love. Aspire. I am so grateful”) (5429RT), while the second leading tweet was written by the leader of the Labour Party in the U.K., Jeremy Corbyn, recalling the fight against fascism (@jemerycorbyn: “This morning, I’m in Portsmouth attending the commemoration of the D-Day landings. The troops that landed on the beaches of France 75 years ago showed unimaginable heroism. Many laid down their lives in the fight against fascism”) (1665RT). While the other top tweets were mainly institutional in origin (e.g., the White House, the U.S. Senate, etc.), one prominent tweet was posted by George Galloway, a Scottish politician, broadcaster and writer in the Respect Party, who took the opportunity to criticise the exclusion of Russia from the ceremonies (@georgegalloway: “Thanks to my dear friend @SameeraKhan for this. It is an offence against history against truth against the memory of the dead to exclude #Russia from #DDay75years”) (766RT). In French, the tweet with the greatest reach came from the literary critic Bernard Pivot, who recalled the international nature of the D-Day landings (@bernardpivot1: “Je donne l’accolade au jeune paysan du Texas et à l’étudiant de Philadelphie qui, demain, il y aura 75 ans, viendront mourir sur une plage de Normandie”) (1358RT). Also noteworthy is the fourth leading tweet with the greatest impact in French, written by Jean Messiha, one of the public faces of France’s National Front, whose community coalesced around the hashtag #6juin1944. Messiha took the opportunity to accuse Emmanuel Macron of “not being French at all” because Macron had omitted passages of a farewell letter written by a young member of the French resistance prior to his execution in 1943 during a reading that Macron gave at the ceremony (JeanMessiha: “#Macron supprime “la France éternelle” de la lettre d’Henri Fertet. Pour ne pas « attiser la haine ni stigmatiser et faire ainsi le jeu du populisme d’extrême-droite », selon les sourates habituelles. Ce président n’a rien de Français ! #6juin1944”) (991RT). Lastly, a prominent tweet came from Edouard Philippe, Prime Minister of France, who recalled the Canadian volunteers who took part in the D-Day landings (@EPhilippePM: “Le 6 juin 1944, les Canadiens qui débarquèrent en Normandie s’étaient tous portés volontaires, au péril de leur vie, pour rétablir la paix. 75 ans plus tard, notre gratitude reste intacte : leur bravoure a contribué à faire de l’Europe un continent libre.”) (345RT). As the tweets show, therefore, there was a plurality of discourses and practices.
In short, the top two languages during the commemoration of the D-Day landings at Normandy featured the discourses of public figures and audiences of significant impact and their tweets were instrumentalised according to the interests present in each context. By contrast, an analysis of the third leading community, which was Spanish, shows that the largest share of tweets was not linked to political parties or public figures, their spread was much more individual in nature and a leading role was played by the portal of political analysis called El Orden Mundial, which also stressed the Soviet role in the conflict (@elOrdenMundial: “Hoy se cumplen 75 años del Día D. ¿Fue tan relevante en la victoria de los aliados? ¿Qué país contribuyó más a la derrota de Alemania en 1945? 1945: URSS (57%) EEUU (20%) Reino Unido (12%) 2015: EEUU (54%) URSS (23%) Reino Unido (18%)”) (179RT) (see Image 1 below).
Image 1. Tweet from @ElOrdenMundial
30-year anniversary of the fall of the Berlin Wall
The ninth of November, which is the date of the commemoration of the fall of the Berlin Wall, is another key date in 2019, marking the 30-year anniversary of the beginning of the process of German reunification in 1989, following 30 years of separation between the German Democratic Republic and the Federal Republic of Germany. The day’s ceremonies were planned and carried out as a visible display of the end of an era that encompassed the Cold War, which had one of its many epicentres in Germany itself. As a result, the online impact was enormous. Specifically, between 4 and 12 November, there were 126,570 tweets (with a reach of as many of 870,593 through retweets), giving the commemoration by far the largest footprint of all the dates in the sample. Once again, the data collection effort has used a combination of web scraping and API monitoring.
For this commemoration, the most influential hashtags on the day were #BerlinWall30 (7.09%), #MuroDeBerlín (2.64%), #BerlinWall (2.51%), #mauerfall30 (1.72%). However, as earlier, the largest share of tweets did not have a hashtag (up to 65.85%). Accordingly, the data collection effort has also covered key words (e.g., Berlin Wall, Berliner Mauer and Muro Berlín). Nonetheless, one of the most important aspects relates to the use of languages. English has been the most frequently used language (45.59%), followed by Spanish (28.22%), French (8.09%), Portuguese (5.3%), German (4.96%) and Italian (3.56%). While the geographical distribution remains unknown, the small minority of tweets in German is certainly striking. Although this fact may be explained partly by the more limited presence of Germans on Twitter based on online user profiles, it is also necessary to examine the use of English for tweets generated in Germany. Indeed, as can be seen below in Figure 6, the diagram shows German included within the English-language community because of their many connections and the permeability of the linguistic boundary between them. This fact reinforces the proposed hypothesis (see Figures 6 and 7 below).
Figure 6. Distribution by language and connections between languages.
Figure 7. Affinity communities within each language group.
As Figure 7 above shows, ideological or party-political subgroups also appear within each language community. The analysis indicates, first, that the English-language community shows a heightened impact of U.S. political parties (Republicans and Democrats), British media outlets and the European Commission and, second, that the event enjoyed widespread dissemination across Latin America and in Spanish media outlets.
In relation to the content of tweets, there are additional phenomena that prove important for this approximation. Beyond any content focusing on the separation of the two Germanies, their reunification and the memory of the period, which tended largely to share in a condemnation of the communist dictatorship, the contemporary political uses of the commemoration stood out. Foremost was the political condemnation of all walls, especially the border policy of Donald Trump. Examples came in tweets from Chris Lu, who had served as an advisor to Barack Obama (@ChrisLu44: “On the 30th anniversary of the fall of the Berlin Wall, Berliners are sending part of the wall to Trump ‘to commemorate the United States’ dedication to building a world without walls.’ Well played.”) (3932RT) and from the news portal Now this News (@Nowthisnews: “These Berliners sent Trump a 2.7-ton piece of the Berlin Wall in protest of his own proposed wall at the U.S.-Mexico border”) (2466RT) (see Image 2 below). Some commentators also took the opportunity to address the situation in China and Hong Kong, for example, in a tweet from the activist Joshua Wong (@joshuawongcf: “30 years ago, the Berlin Wall fell, and we hope that in the future, the great firewall of China will also fall. We hope for more and more people who live in mainland China to recognise the importance of universal values and freedom”) (4848RT) or they commented on political unrest in Chile and Bolivia, for example, in a tweet from the far-right Chilean leader José Antonio Kast (@joseantoniokast: “Hace 30 años, la libertad hizo añicos el Muro de Berlín y la derrota de la sangrienta ideología comunista era evidente. Hoy en Chile, los mismos comunistas y la izquierda radical nos quieren”) (2256RT). The last two examples had a strong impact as a result of their presentist use of the commemorations, drawing parallels with other purportedly authoritarian regimes or with restrictive policies aimed at immigrants, and they generated a high degree of social engagement and condemnation, while also standing in stark contrast to the official or institutional line.
Image 2. Tweet from @Nowthisnews
European day of remembrance for the victims of all totalitarian and authoritarian regimes
The day of remembrance for the victims of all totalitarian and authoritarian regimes is celebrated on 23 August in memory of the Molotov-Ribbentrop Pact between the Soviet Union and Nazi Germany in 1939. Highly controversial, the date is the subject of a recurring debate over the political justifications that lend it legitimacy and that equate communism with Nazism. Despite the political furor, however, the commemoration has had a very limited impact on social media. Specifically, between 20 and 26 August 2019, Twitter registered a total of only 4,527 tweets (with a reach of as many as 23,054 through retweets), which is the smallest digital footprint of any commemoration in the sample. Once again, the data collection effort has used a combination of web scraping and API monitoring.
The most frequently used hashtags were #BalticWay30 (26.38%), #BlackRibbonDay (10.21%), #TheHongKongWay (5.53%) and #80WW2 (3.01%). In other words, the commemoration of the Baltic Way, a human chain formed on 23 August 1989 to mark the independence of the Baltic countries, exceeded the ceremonies hosted by the European Parliament. It also generated a homologous, symmetrical trend in Asia focusing on the conflict in Hong Kong. Another example of the unique nature of the event lies in the very small amount of content that did not have a hashtag (only 7%).
The most common language of use on the day was again English (68.02), followed by Polish (7.72%), Japanese (5.43%), Spanish (3.6%), Greek (2.89%), German (1.53%) and French (1.18%). In other words, language diversification has resulted from decentralisation and from the fact that the memory related to the commemoration is not situated solely in Western Europe. In these trends, it has also been possible to identify different groups located in the Baltic countries and Hong Kong as well as a Japanese segment revolving around the TV channel jptime.tv (see Figures 8, 9 and 10 below).
Figure 8. Distribution by language.
Figure 9. Distribution by community (notice when comparing the diagram here with the one in Figure 8 that the Lithuanian and Latvian communities tweet in English).
Figure 10. Distribution by hashtag. (The most frequently used hashtag was #BalticWay30, which was related to the celebration of #BlackRibbonDay, but not the same. Next appeared #TheHongKongWay and #HongKongWay, which were also related to #BalticWay30. The Japanese hashtag 香港之路 (in English, Hong Kong Way) was likewise related to #HongKongWay, while #80WW2 referred to the anniversary of the Molotov-Ribbentrop Pact. The remaining hashtags reflect variations of lower-case and upper-case letters. Ultimately, Poland and the United States/Canada were the places where the day had the greatest impact.)
At the content level, the most prominent public figures and institutions in Western Europe included accounts that belonged to the European Parliament in France (@Europarl_F: “En ce jour du souvenir, nos pensées vont aux victimes du nazisme et du stalinisme. Le pass én’est jamais vraiment mort et nous n’oublions pas l’époque sombre du totalitarisme. #BlackRibbonDay”) (22RT) and to the London mayor Sadiq Khan (@MayorofLondon: “Today marks the European Day of Remembrance for Victims of Stalinism and Nazism. Now more than ever we must show our commitment to fighting extremism, authoritarianism and intolerance in all its forms.”) (752RT). While the online community in the Polish language hardly made any reference to the Baltic Way, a tweet from Poland’s Foreign Office did have a major impact (@MSZ_RP: “23.08.1939: Rzesza Niemiecka i ZSRS zawarły pakt o nieagresji podpisany przez J. von Ribbentropa i W. Mołotowa. Umowa zawierała tajny zapis o podziale terytoriów suwerennych państw i We wrześniu oba państwa napadły na rozpoczynając IIWŚ #80WW2 #BlackRibbonDay”) (370RT). Trending topics on the Baltic Way, by contrast, were much more numerous, reflecting the participation of political figures in Western and Eastern Europe, including the European Commission (@EU_Commission: “Today we mark the 30th anniversary of the Baltic Way. Around two million people from Estonia, Latvia and Lithuania formed a 600km long human chain, demonstrating their unity in their efforts towards freedom. They helped overcome divisions and unify Europe. #BalticWay30”) (1881RT [compared to 406 for a tweet sent hours later mentioning the victims of authoritarian regimes]) and Baiba Braze, Latvia’s ambassador to the United Kingdom (@BaibaBraze: “30years ago with family, friends and other students I stood in the 2 million strong 600km long human chain for freedom & future of #Estonia #Latvia #Lithuania, against the NaziSoviet division of Europe. We won, and now treasure our European way. #BalticWay30”) (389RT). By contrast, the community revolving around the Asian phenomenon was largely situated in Japan and China, with the biggest share of tweets appearing in Japanese.
In conclusion, it can be confirmed that the day in memory of the victims of totalitarian regimes, like Europe Day, enjoys very little engagement on Twitter, though it must also be added that its political justification has not been without controversy. It is, therefore, unsurprising that the celebration of 23 August was eclipsed by other commemorations that enjoy more wide-reaching acceptance and a public memory that can moreover generate discursive replies around the world, as in the case of Hong Kong.
Commemorations in Spain
Day of the liberation of Mauthausen concentration camp
The day of the liberation of Mauthausen concentration camp has been included in the Spanish section because of its impact on Spain’s political and memorial agenda in 2019, based on the government’s participation in the tribute and the cabinet’s approval and recognition of 5 May as a day to remember the Spanish victims of Nazism. The event also served as an indirect recognition of exiled republicans and their memory, a practice that has never become well-established in democratic Spain. However, the analysis of the commemoration must remain necessarily superficial, because the only available information at present consists of global metadata. There are no visual displays or analyses by community, hashtag, language of use or specific hour of the day. The pertinent data collection effort took place in the initial phases of the project, when the methodologies and approaches were still being refined. Based on the original tweets and subsequent retweets, however, the day had a reach of 77,993 interactions. In other words, the figure shows a high degree of impact nationwide, which comes as a surprise in light of limited support for the commemoration at the institutional level. Based on an analysis of the most prominent content (i.e., tweets with the greatest number of retweets), the most plausible explanation is primarily political and contextual in nature.
The Catalan political conflict erupted directly in the midst of the commemoration events, when the Director-General of Memory within the Catalan government, Gemma Domènech, took the opportunity to recall the Catalan political prisoners currently behind bars, especially Raúl Romeva, who had been serving as minister in charge of overseeing the Catalan government’s memory policies before his sacking and incarceration. Domènech’s remarks prompted the Spanish minister of justice, Dolores Delgado, to stand immediately and walk out of the ceremony.
One indicator of the high presence of these events in the online footprint of the day is the language itself, given that 21,970 of the total interactions noted above occurred in Catalan. This represents a high and very significant proportion. Similarly, out of the total, 19,718 contained the word Romeva, 10,620 contained Catalunya (the Catalan spelling of Catalonia) and 17,458 contained Generalitat (the Catalan name of the government of Catalonia). Another striking fact relates to the profiles that proved very active in tweets that were representative of both sides of the conflict. Two examples were a tweet from Esquerra Republicana politician Gabriel Rufian @gabrielrufian, who omitted facts and criticised the Spanish government for its failure to honour the memory of the Second Spanish Republic (“Hoy que es noticia lo de Delgado en Mauthausen: 1) Presidentes españoles se han negado a ir por las banderas republicanas del acto. 2) Todos los gobiernos españoles se han negado a asumir responsabilidades por las deportaciones del franquismo. Por si no te lo están explicando.) and the representative of Ciudadanos Juan Carlos Girauta (@GirautaOficial: “Han ido a Mauthausen a rendir homenaje al antisemita Romeva, que consagra en la Eurocámara más esfuerzos que nadie para impulsar el movimiento BDS. El nacionalismo catalán es basura”). Other prominent political profiles were @ahorapodemos (“El 5 de mayo de 1945 fue liberado el campo de concentración de Mauthausen donde deportaron 10.000 republicanos españoles, condenados a muerte por el franquismo y al olvido por todos los Gobiernos posteriores aquellos que sobrevivieron.”) and the president of the Catalan government Quim Torra (@QuimTorraiPla: “Moltes gràcies @gemmadomenech i @catmemoria per la vostra dignitat, i per recordar al Conseller @raulromeva , pres polític d’Espanya. Gràcies per portar la veu del Govern de Catalunya en l’acte d’homenatge a tots els catalans i republicans de l’Estat asassinats a Mauthausen”). A few days later, the historian Rosa Toran, speaking on behalf of the Junta del Amical de Mauthausen and other concentration camps, published an article in which she lamented the fact that the many months of advance preparations and the emotional rapport developed across multiple generations during the four days of events (which had drawn on the participation of 300 people and 30 institutes from across Spain) should be eclipsed and tossed in the rubbish bin because of what she called, in essence, “controversies of little importance”.
To provide more in-depth analyses, however, more specific data would be required. At present, it is only possible to confirm the high volume of data generated by the day and the political unrest arising from the Catalan conflict that pervaded the commemoration.
Commemoration of the Second Spanish Republic
The date of 14 April is an unofficial commemoration in the Spanish calendar. It involves no public acts of tribute or institutional memory at the level of the Spanish state. However, the date does have a huge symbolic value for a large part of the population and has become a grassroots commemoration in remembrance of the freedom that was cut short by the outbreak of the Spanish Civil War and the ensuing Franco regime. It has also turned into a controversial date marked by conflict, which, to a large extent, replicates the dynamics of clashes between the left and the right and conveys the inadequate political management that has surrounded the period in question. Accordingly, it has been decided to include an analysis of the commemoration’s online footprint in order to examine the potential of an unofficial date that remains alive in memory and faces constant resignification. To grasp its reach, the first indicator at the level of the volume of data shows that between 12 and 17 April, with 14 April as the middle date, there were 11,180 tweets generated (with a reach of as many as 86,714 through retweets). This appears to be a significantly high presence when compared to other commemorations at the European level, particularly taking into account that the phenomenon mainly takes place only in Spain. It is also necessary to bear in mind, however, that the Spanish-speaking community is generally one of the most active overall on Twitter, as the sociological profile attests. Once again, the data collection effort has adopted a combination of web scraping and API monitoring as the chief methodology.
The most influential hashtags on the day were #14deAbril (55.51%), #14DeAbril (7.29%), #14deabril (5.69%) and #VivaLaRepública (2.12%), that is, variations of the same hashtag combining upper-case and lower-case letters in different formulations. Also, the tweets without a hashtag were very limited (5.04%), reflecting a trend toward the homogenisation of communities and the small number of conversations not situated on one axis or another. Also, it is relevant to point out that no language data are available for the day. As a result, it has not been possible to calculate the level of content generated in languages other than Spanish, such as Catalan, Basque and Galician. When affinity communities are identified, however, the results show a high level of political polarisation that is without precedent on the other dates in the sample. Indeed, as Figure 11 below shows, there are two perfectly identified blocs that are separated from one another and have very few connections between them: first, the bloc on the left, which is a majority that supports the commemoration the Second Spanish Republic; and second, the bloc on the right, which opposes the commemoration. At the same time, each of the two blocs can be divided into subgroups. The bloc on the left is associated with political parties like Unidas Podemos and Izquierda Unida and with republican groups, while the bloc on the right prominently features far-right party Vox and other neoliberals. Overall, the leading profiles corresponded to political parties. On the left, they included @ahorapodemos (“Republicanismo significa que cualquier ciudadano va a tener garantizado su derecho a un empleo y una pensión digna, a que su familia pueda ser atendida en los mejores hospitales y a estudiar en las mejores universidades. @Pablo_Iglesias_”) and @Iunida (“izando la tricolor en Eibar. #14deabril”) (see Image 3 below). On the right, they included @vox_es (“El 14 de abril lo único que celebramos en VOX es el cumpleaños de nuestro presidente. Felicidades @Santi_ABASCAL por tus 43 #14deAbril #YoNoMeRindo”). Other prominent user profiles were @ecorepublicano (“Feliz Día de la República #14deAbril #VivaLaRepública”), @MiecolesRepub1 and @joveneuropeon (“cosas que posiblemente no te han explicado de la Segunda República: censura de prensa, constitución aprobada sin referedúm, la Ley de Vagos y Maleantes… un poco de esa memoria histórica que no le gusta a la izquierda” #14deabril).
Figure 11. Distribution by affinity communities.
For this analysis, it has been possible for the first time to include an approximation of the distribution by gender. While this indicator is not present in Twitter profiles, it can be extracted indirectly by analysing the names that appear in profiles. To this end, the classification by gender has been carried out using the list of most common names compiled by the Spanish National Statistical Office (INI). Gender has been identified in 42.13% of profiles. Notably, the proportion of women (20.13%) compared to men (22.01%) is slightly higher than the percentage of women in existing Twitter accounts according to the Spanish Centre for Sociological Research (CIS), which has identified 30% women and 38% men.
Image 3. Tweet from @iunida
In conclusion, the commemoration of the Second Spanish Republic has shifted from an underground remembrance during the long period of Francoism to a living memory now, both on social networks and in society itself, even though it is not officially recognised at the national level. For a portion of the population, 14 April continues to represent an episode in Spanish history that is bound up with values like republicanism, freedom and democracy. Also evident, however, is an attempt to discredit the Second Spanish Republic by the far right, which seeks to counter this positive view with a negative account of the period that preceded the outbreak of civil war in 1936. As a result, the date can also be viewed as a target for latent political conflict and stark polarisation driven by political parties at both extremes.
Conclusions and challenges for the future
The project Online Memories and its findings offer a new outlook with major potential for use in sociological analysis and the examination of online uses in order to discern, on one hand, the depth of the social influence of commemorations and other key dates on the calendar and, on the other hand, the appropriations and instrumentalisations of large organisations and political parties, based on the trending topics generated online. These elements are important in a time of rising far-right populism and increased fake news that together contribute to the mutation of memory practices – an increasingly familiar feature – and to the spread of disinformation. At the same time, there is also an opportunity for spontaneous transformations initiated from the lower strata of society, which is evident in the start of new forms of commemoration that are much more permeable and cooperative in their construction. As a result, social media can have both negative and positive potential for practices and uses related to the past, above all among both the younger and middle-aged strata of the population. Not only for this reason but also because of the reach that online phenomena may attain in the years to come, it is important to carry on with these sorts of analysis in the future, expanding the dates, commemorations, and types of data that can be extracted. Examples include improving the identification of communities and micro-communities in order to differentiate political and institutional uses from more grassroots content; looking at gender variables; better situating geographical positions; breaking down sociological information by region, and providing more detailed information on the countries of Eastern Europe, which are not well represented at present. At the same time, it will also be valuable to make stronger comparisons between official holidays and unofficial ones that have a certain level of impact on social media. The data relating to 14 April clearly point to the potential for an analysis of these categories.
A further aim will be to expand on the findings in this report with more detailed comparisons, given that a large amount of data has necessitated a process of synthesis and prioritisation. Nonetheless, it is possible to venture some conclusions that have been emerging throughout the analysis and that will prove significant for future research.
The first point concerns the phenomenon of the language barrier, which has appeared repeatedly on the preceding pages. Despite the growing trend toward homogenisation around the use of English, Romance-language communities mainly are persisting in the use of their local languages. In this respect, Spanish is typically the language with the second greatest reach, although there is also a presence of French and Italian. That said, there remains a need to expand the geographical scope to refine these conclusions in the case of other regions. Nonetheless, the interesting aspect of this language barrier lies in the low amount of connections between language communities, which were very limited in nearly every commemoration in the sample (with the exception of the centenary of the end of the First World War, when French and English tended to relate to one another as a matter of historical unity). To grasp this reality, it is also necessary to take into account that Twitter has a simultaneous translation function. This is, therefore, a cultural issue too, since an understanding of English is quite widespread and yet online connections arise between countries only in the case of bilateral historical processes. In the future, it will be necessary to deepen this analysis and look at connections by affinity community and axes of polarisation.
The second point is that it is possible to identify the level of comparative impact of the various commemorations in order to draw a conclusion, as yet partial, on their social and online reach (see Graph 4 below).
Graph 4. Prepared by the author.
In some cases, the differences among commemorations are extreme. Such differences are far greater for one-off dates linked to non-structural events and anniversaries, such as the 30-year anniversary of the fall of the Berlin Wall and the centenary of the end of the First World War, the two dates of greatest impact in 2019. Other dates, such as the anniversary of the D-Day landings at Normandy, Europe Day and Holocaust Remembrance Day, by contrast, had a similar impact to one another and show some degree of stability. To the contrary, the Remembrance Day for Victims of Totalitarianism, which also included the celebration of the #BalticWay, had a very small impact that must be better understood. Certainly, its under-representation may be accounted for partly by the limited engagement generated by the date in Western Europe and partly by the lack of better research on the technological and online profile of countries in Eastern Europe, where the date had a greater impact. Lastly, the specifically Spanish commemorations show a high degree of impact, particularly bearing in mind that they were limited to a single country. Their impact can be explained by the high level of activity among the Spanish and Spanish-speaking community on Twitter.
Ultimately, it will be necessary to enhance the method for processing content, separating tweets with greater reach by affinity community and not only by language community, in order to identify the trending topics and agents that have the greatest influence within the various groups. This level of processing will be essential to an analysis of political use on Twitter and an exploration of how fake news and grassroots resignifications are generated.
 Taken from Omnicore and Statista.
 The Italian language community stands slightly apart from the others and it makes fewer connections. This phenomenon is partly due to the fact that the hashtag #11Novembre was also used for a referendum on metropolitan transport held on the same day in the city of Rome. For this reason, the Italian community might also contain false positives and appear over-represented.
 It can be confirmed, however, that the controversy over Pétain had a major impact and that, at the same time, the remembrance of the Armenian genocide also had a considerable presence on the date (e.g., @N_Kechichian: “L’affaire #Pétain a réussi à polluer les debats et a empêché l’évocation du #Genocide des #armeniens de la première guerre mondiale dans cette commémoration de l’#Armistice de 1918”  and @BLEUKLEINN: “Je vous rappelle Monsieur @EmmanuelMacron que Philippe Pétain a été responsable de la déportation de 75000 juifs français . Moi Française juive petite fille d’un juif décoré en 14/18 puis banni parce que juif sous Pétain je n’oublierai pas votre insulte à l égard de nos familles” [900 RT]).