Digital Media

Existe uma crescente procura e interesse em  compreender  fenómenos  culturais  e sociais  através  das  redes  sociais
online (Facebook,  Twitter,  etc.).    Apesar  dos avanços tecnológicos  e  da  enorme  lista de  ferramentas  computacionais,  não  há evidências que comprovem sobre o uso,
por  exemplo,  de  métodos  digitais  (Rogers,  2013)  para estudos  sobre  o  Facebook em Portugal.  O grau de incerteza
e pouco conhecimento sobre o problema citado leva-nos a propor uma revisão sistemática (Petticrew; Roberts, 2006; Kit-
chenham, 2007) em busca de provas concretas acerca dos métodos adotados e utilizados nos últimos três anos em Portu-
gal.  Para esse objetivo foram selecionadas universidades,  congressos e revistas científicas de referência no país.  O ob-
jetivo da revisão foi promover um background
para novas práticas de investigação no campo das ciências comunicacionais.  O artigo foi dividido em três partes: i) primeiro uma breve introdução sobre os novos métodos e respetivas opor-
tunidades  para  o  estudo  das  redes  sociais;  ii)  em  seguida  são  apresentados  os parâmetros adotados nesta revisão siste-
mática; iii) por fim, apresentam-se os resultados,  que revelam os principais métodos usados para estudos sobre o Face-
book em Portugal.

Palavras-chave: revisão sistemática, Facebook, métodos digitais, métodos computacionais

Interactivity is introduced here not only as a key concept for communication, but also as one useful
perspective for studying social media, especially Facebook (FB) pages and groups. Some words are
inherent to the concept of interactivity such as connection, meaning making, continuum and community
orientation – taking into account the individual experience and subjective perceptions. In line with these
basic features and considering interactivity as essential measure (or cause) of the social dynamics of
group communication (Rafaeli and Sudweeks, 1994), its concept and actual realization seem to be closely
related to the interaction on a FB page or group. In this context, we wonder how can one visualise and
measure interactivity. Digital Methods propose a unique form to comprehend society and culture through
methods embedded within the first language of online social networks; codes. In order to find clear
answers and endeavouring to visualise and to measure interactivity, this paper proposes a study of
Brazilian’s 2013 protests on Facebook by using Digital Methods. We intend to discuss key concepts of
interactivity and its relation with FB, as well as, a short explanation of Brazilian’s 2013 Protests; then, we
are going to give a brief introduction of Digital Methods – presenting pivotal ideas of Network Theory
and Social Network Analysis; and, finally, we aim to visualise and measure interactivity through graphs
and graphics.

[Paper presented at From Multitude to Crowds in Social Movements – publics, gatherings, networks and media in the 21th century Lisbon, 26 and 27 January 2015 An International Conference hosted by the Catholic University of Portugal in a CECC/CECL co-organization]

A Internet, em especial por intermédio das plataformas de redes sociais, tem cada vez mais desempenhado um papel como meio de organização, divulgação e exposição das mobilizações sociais ao redor do mundo. Tem igualmente sido alvo de interesse em estudos académicos, embora a maioria destes adote abordagens fundamentadas em práticas tradicionais. No Brasil, os protestos de junho de 2013 geraram, de certa forma, uma mudança de postura em massa por parte dos brasileiros perante a participação política. Dois anos após as manifestações que marcaram a história do país, os brasileiros foram novamente para as ruas (nos dias 15 de março, 12 de abril, 16 de agosto e 13 de dezembro de 2015) descontentes com a gestão presidencial de Dilma Russeff, com os escândalos de corrupção e com o baixo crescimento económico do país. Neste contexto, propomos um estudo experimental a partir dos Métodos Digitais procurando compreender o 15 de Março através das redes sociais – mais concretamente o Facebook – confrontando as páginas de mobilização direta dos protestos e com uma de teor mais cómico. Para tal, dividimos o artigo em quatro partes: i) o duo movimentos sociais e redes sociais e respetivas relações, debates e percepções de estudo; ii) os protestos no Brasil numa breve contextualização das Jornadas de Junho (2013) e dos movimentos Anti-Dilma em 2015; iii) o guião metodológico; por fim, iv) análise dos resultados. Palavras-chave: 15 de Março, Protestos Anti-Dilma, Brasil, Métodos Digitais, Facebook. 1. O duo movimentos sociais e redes sociais A ligação entre movimentos sociais e plataformas de redes sociais assinala uma nova forma de mobilização social; mais rápida e eficaz na sua capacidade de reunir, divulgar e denunciar. A expressão massiva (ou pontual) dos valores coletivos em prol de interesses comuns faz destas redes, mais do que ferramentas de agrupamento e organização, novos meios para a participação cívica e política. Estudar as mobilizações no Facebook ou Twitter, por exemplo, conduz-nos à identificação de narrativas coletivas (seus principais atores e tópicos de discussão) e permite-nos compreender e analisar a opinião e os atos políticos sobre um tema específico. Na medida em que dão voz, fazem ruído e geram mudanças, os SRS 2 são uma nova forma de fazer política. Entendemos o duo movimentos sociais e SRS a partir de três pressuposições: i) os SRS são parte integrante dos movimentos sociais, ferramentas que os moldam e desenvolvem algumas de suas capacidades (Tufekci, 2014); ii) podem também ser objecto de crítica, de acordo com Rogers (Digital Methods Initiative, 2015, 1 Junho) e podem constituir uma forma alternativa 1 Com um especial agradecimento a João Fonseca pelo imprescindível auxílio com gráficos no Excel. 2 Ainda que consideremos preferível o termo «plataforma» (cf. Helmond, 2015), mantemos ocasionalmente a sigla mais comum, SRS («sites de redes sociais»).

Hashtag engagement research is proposed through four main perspectives (media item, user, visual and grammars of hashtag), drawing the rise of political polarization in Brazil as a case study (pro-impeachment and anti-coup protests in March 2016). After exploring three perspectives, our general key findings: 1) the awareness of Instagram as a platform to ground studies on imagery of polarised political engagement. 2) That there is substantial difference between the panorama of an issue if we observe the most liked/commented/shared media items and the panorama that can be observed in the less popular ones. Thus, it is insufficient to focus the analysis on the 'top' items and users. In what regards key findings from each perspective to visualise hashtag engagement: i) Media item perspective: The visual culture of the dominant voices presented similar composition, structures and adoption of visual elements in both protests, for instance, composed by public figures (pro-impeachment: actors or actresses, TV presenters, stand-up comedians, businesswoman; anti-coup: politicians, artists, LGBT and feminists activists, Independent Media) with few publications but accounting the most engaged media items, and the organisers of the protests or activists and non-official campaign accounts, these actors will probably not appear in the top ten most engaged lists, they have more publications sustaining influence and visibility along the day of the protests. It was curious to note the presence of the account petscharm in the dominant voice list of pro-impeachment protests; this profile depicted different breed of dogs wearing the official uniform of Brazil's football team or Brazilian flag in all publications. However, and despite common characteristics, a close look at dominant voices revealed also the predominance of visual elements: green and yellow, selfies, Brazil's football garment, and dogs for pro-impeachment; and red, political slogans, crowds, strong critique on Globo television. And lastly, along the anti-coup protests brazilians were more engaged with publications made by public figures, whereas, in pro-impeachment protests, publications made by the organisers of the protest and activists or non-official campaign accounts drew more attention from protesters (see table below). 2 ii) User perspective: The text extraction of the captions of dominant and ordinary voices and its visualisation by clusters of tags, terms and political position show that there are a lot of common discourse among the groups. But looking further, we could see that dominant voices scored more in specificity and frequency when the post was about the locations of the protest. This suggests that dominance itself (the capacity of attract and maintain the attention/engagement) may not necessarily be related to the caption or hashtag content, but to pragmatic information with regards to protest points and locations, or their organising power. When it comes to comparing anti-coup (left-wing) and pro-impeachment (right-wing) user perspectives, all the analysis showed the pro-impeachment groups larger and well-connected, although they were fewer. They also included the protest-location group and, therefore suggesting again that the dominant voices in the pro-impeachment group were more connected and (through the connections and the protest-locations) more tightly and therefore potentially better organised. Among the ordinary voices, we could see more connections between the two main political groups (pro-impeachment and anti-coup). The mutual space seems to be linked to the aim of gathering and engaging people to a cause, stressing the collective dimension and setting a common horizon and agenda. iii) Visual perspective: Analysis of the visual content of pictures related to each protest (pro-impeachment and anti-coup) revealed a common general structure of the images of each protest. Three broad categories emerged of the bipartite image-label network, making it largely a triangular network with vertices related to: open-shots of the crowd, selfies and close-up portraits, and graphics (invitations to the protest, banners etc.). Also, both protests had a smaller label cluster related to 'foods', specifically, related to two types of food which are pejoratively attributed to each side of the dispute. As differences between protests, apart from labels related to the colors worn by partisans of each group, there were sub-clusters which reveal the manifestation of some stereotypes of each group. In the anti-coup group there is a subcluster in the portrait/selfie cluster related to beard and facial hair. In the pro-impeachment, we found a subcluster related to sunglasses and a small autonomous cluster related to dogs (most of which depicted wearing the yellow uniform of Brazilian football team). While many questions related to this perspective were left untouched due to operational issues in the research process, the research showed a potential for automatic annotation in revealing broad descriptions of a visual dataset. Some aspect to later pursue include the question of variability of these descriptions related to particular hashtags and between dominant and ordinary voices. (see final presentation)

Analysis of social media using digital methods is a flourishing
approach. However, the relatively easy availability of data collected
via platform application programming interfaces has arguably led
to the predominance of single-platform research of social media.
Such research has also privileged the role of text in social media
analysis, as a form of data that is more readily gathered and
searchable than images. In this paper, we challenge both of these
prevailing forms of social media research by outlining a
methodology for visual cross-platform analysis (VCPA), defined as
the study of still and moving images across two or more social
media platforms. Our argument contains three steps. First, we
argue that cross-platform analysis addresses a gap in research
methods in that it acknowledges the interplay between a social
phenomenon under investigation and the medium within which it
is being researched, thus illuminating the different affordances and
cultures of web platforms. Second, we build on the literature on
multimodal communication and platform vernacular to provide a
rationale for incorporating the visual into cross-platform analysis.
Third, we reflect on an experimental cross-platform analysis of
images within social media posts (n = 471,033) used to
communicate climate change to advance different modes of
macro- and meso-levels of analysis that are natively visual: imagetext
networks, image plots and composite images. We conclude by
assessing the research pathways opened up by VCPA, delineating
potential contributions to empirical research and theory and the
potential impact on practitioners of social media communication.

Created in the end of 2015 and currently approaching the mark of 200 thousand followers, the Facebook page " Os Truques da Imprensa Portuguesa " [The Tricks of the Portuguese Press] posts critical remarks on the news of national media and their supposed editorial criteria. The page has often generated heated debates on the platform, either being praised for its role as a watchdog, or discredited as allegedly serving as the spokesperson for a hidden political agenda. Until recently, the anonymity of their writers was one of the arguments for this accusation, but when the identity of one of the admins was disclosed, promptly followed by the self-disclosure of both, this premise was rebutted. This paper is focused on the reactions to the post on July 9th 2017 in which the admins revealed their names. Our goal is to evaluate the debate generated by this post, in particular concerning the polarization of positions and arguments among those that engaged with the post. To this end, we conduct a mostly exploratory research supported by visual network analysis and textual analysis. We first provide a global characterization of the page, and subsequently present some insights on the revealing post advanced by visual network analysis. Lastly, we look into textual content considering the " global mindset " of the discussion (topics, subtopics, and positionings for and against the page or the post), looking for eventual differences between the subsets of more active and less active commenters. We conclude that the engagement with the post (especially comments) was less polarized than expected, although with one notable exception, and that the relevance of the post rested more in its potential to fuel the debate around contemporary changes and practices in journalism and its multiple dimensions than in the positionings themselves.

The workshop “Networks, Hashtags, Memes: A quali-quantitative approach for exploring social media engagement” ​with Janna Joceli Omena and Elena Pilipets addresses the issues of circulation, appropriation and transformation of user-generated content in networked environments of social media platforms. To do so, it combines digital methods of data extraction, network visualisation and engagement analysis with qualitative approaches of situational mapping and internet ethnography focusing on searchable visual content.

1. Studying Hashtag Engagement through Digital Networks (and methods!)|  Janna Joceli Omena

The first part of the workshop introduces a medium-research perspective with focus on digital networks and the study of circulation, appropriation and transformation of hashtagging activity. Undertaken as socio-technical representational schemes of trackable-retrievable actions, digital networks offer ways of understanding social and cultural phenomena. Rather than relying on the common use of statistics, the workshop brings a theoretical, practical and technical reflection that combines the knowledge of platform grammatisation with the praxis of data capture and data analysis. With the aim of exploring hashtagging engagement through the visual affordances of digital networks, we will first discuss a framework for interpreting digital networks. This will be followed by a step-by-step walk through a case study on the circulation of image content about Brazilian presidential elections in 2018.

2. Situating Internet Memes as Mediators & Techno-Social Multiplicities | Elena Pilipets

In the second part of the workshop, we will experiment with some qualitative approaches to contextualising/analysing/conceptualising the circulation of user-generated content (viral images, memes, gifs) on social media platforms. Our focus will be on addressing the networked formations of internet memes, which are ephemeral and contextually situated as much as they are spreadable/durational. Beyond looking at how memetic adaptations reproduce and shift networked relations of visibility/affectivity, we will explore the highly ambiguous alliances of platform recommendation and subcultural stance in users’ practices of content appropriation. With the emphasis on understanding the digital circuits of visual social exchange through a combination of mediated thick descriptions and situational mapping​, we will discuss the concepts of networked attention economy and social media affordances in the context of analysing memetic engagement.