Studies of political polarization in social media demonstrate mixed evidence for whether discussions necessarily evolve into left and right ideological echo chambers. Recent research shows that, for political and issue-based discussions, patterns of user clusterization may differ significantly, but that cross-cultural evidence of the polarization of users on certain issues is close to non-existent. Furthermore, most of the studies developed network proxies to detect users’ grouping, rarely taking into account the content of the Tweets themselves. Our contribution to this scholarly discussion is founded upon the detection of polarization based on attitudes towards political actors expressed by users in Germany, the USA and Russia within discussions on inter-ethnic conflicts. For this exploratory study, we develop a mixed-method approach to detecting user grouping that includes: crawling for data collection; expert coding of Tweets; user clusterization based on user attitudes; construction of word frequency vocabularies; and graph visualization. Our results show that, in all the three cases, the groups detected are far from being conventionally left or right, but rather that their views combine anti-institutionalism, nationalism, and pro- and anti-minority views in varying degrees. In addition to this, more than two threads of political debate may co-exist in the same discussion. Thus, we show that the debate that sees Twitter as either a platform of ‘echo chambering’ or ‘opinion crossroads’ may be misleading. In our opinion, the role of local political context in shaping (and explaining) user clusterization should not be under-estimated.

The paper is dedicated to solving the problem of optimal text classification in the area of automated detection of typology of texts. In conventional approaches to topicality-based text classification (including topic modeling), the number of clusters is to be set up by the scholar, and the optimal number of clusters, as well as the quality of the model that designates proximity of texts to each other, remain unresolved questions. We propose a novel approach to the automated definition of the optimal number of clusters that also incorporates an assessment of word proximity of texts, combined with text encoding model that is based on the system of sentence embeddings. Our approach combines Universal Sentence Encoder (USE) data pre-processing, agglomerative hierarchical clustering by Ward’s method, and the Markov stopping moment for optimal clustering. The preferred number of clusters is determined based on the “e-2” hypothesis. We set up an experiment on two datasets of real-world labeled data: News20 and BBC. The proposed model is tested against more traditional text representation methods, like bag-of-words and word2vec, to show that it provides a much better-resulting quality than the baseline DBSCAN and OPTICS models with different encoding methods. We use three quality metrics to demonstrate that clustering quality does not drop when the number of clusters grows. Thus, we get close to the convergence of text clustering and text classification.

Today, aggressive verbal behavior is generally perceived as a threat to integrity and democratic quality of public discussions, including those online. However, we argue that, in more restrictive political regimes, communicative aggression may play constructive roles in both discussion dynamics and empowerment of political groups. This might be especially true for restrictive political and legal environments like Russia, where obscene speech is prohibited by law in registered media and the political environment does not give much space for voicing discontent. Taking Russian YouTube as an example, we explore the roles of two under-researched types of communicative aggression—obscene speech and politically motivated hate speech—within the publics of video commenters. For that, we use the case of the Moscow protests of 2019 against non-admission of independent and oppositional candidates to run for the Moscow city parliament. The sample of over 77,000 comments for 13 videos of more than 100,000 views has undergone pre-processing and vocabulary-based detection of aggression. To assess the impact of hate speech upon the dynamics of the discussions, we have used Granger tests and assessment of discussion histograms; we have also assessed the selected groups of posts in an exploratory manner. Our findings demonstrate that communicative aggression helps to express immediate support and solidarity. It also contextualizes the criticism towards both the authorities and regime challengers, as well as demarcates the counter-public.

Ive summarization is a technique that allows for extracting condensed meanings from long texts, with a variety of potential practical applications. Nonetheless, today’s abstractive summarization research is limited to testing the models on various types of data, which brings only marginal improvements and does not lead to massive practical employment of the method. In particular, abstractive summarization is not used for social media research, where it would be very useful for opinion and topic mining due to the complications that social media data create for other methods of textual analysis. Of all social media, Reddit is most frequently used for testing new neural models of text summarization on large-scale datasets in English, without further testing on real-world smaller-size data in various languages or from various other platforms. Moreover, for social media, summarizing pools of texts (one-author posts, comment threads, discussion cascades, etc.) may bring crucial results relevant for social studies, which have not yet been tested. However, the existing methods of abstractive summarization are not fine-tuned for social media data and have next-to-never been applied to data from platforms beyond Reddit, nor for comments or non-English user texts. We address these research gaps by fine-tuning the newest Transformer-based neural network models LongFormer and T5 and testing them against BART, and on real-world data from Reddit, with improvements of up to 2%. Then, we apply the best model (fine-tuned T5) to pools of comments from Reddit and assess the similarity of post and comment summarizations. Further, to overcome the 500-token limitation of T5 for analyzing social media pools that are usually bigger, we apply LongFormer Large and T5 Large to pools of tweets from a large-scale discussion on the Charlie Hebdo massacre in three languages and prove that pool summarizations may be used for detecting micro-shifts in agendas of networked discussions. Our results show, however, that additional learning is definitely needed for German and French, as the results for these languages are non-satisfactory, and more fine-tuning is needed even in English for Twitter data. Thus, we show that a ‘one-for-all’ neural-network summarization model is still impossible to reach, while fine-tuning for platform affordances works well. We also show that fine-tuned T5 works best for small-scale social media data, but LongFormer is helpful for larger-scale pool summarizations.

The community-based structure of communication on social networking sites has long been a focus of scholarly attention. However, the problem of discovery and description of hidden communities, including defining the proper level of user aggregation, remains an important problem not yet resolved. Studies of online communities have clear social implications, as they allow for assessment of preference-based user grouping and the detection of socially hazardous groups. The aim of this study is to comparatively assess the algorithms that effectively analyze large user networks and extract hidden user communities from them. The results we have obtained show the most suitable algorithms for Twitter datasets of different volumes (dozen thousands, hundred thousands, and millions of tweets). We show that the Infomap and Leiden algorithms provide for the best results overall, and we advise testing a combination of these algorithms for detecting discursive communities based on user traits or views. We also show that the generalized K-means algorithm does not apply to big datasets, while a range of other algorithms tend to prioritize the detection of just one big community instead of many that would mirror the reality better. For isolating overlapping communities, the GANXiS algorithm should be used, while OSLOM is not advised.