More and more digital data is being generated every day, and more and more social science
analyses are using Twitter, Instagram, or Facebook data. Many international and national studies
have already explored the social science opportunities and dilemmas raised by the phenomenon
of ‚big data’ - but the issue of ‚access to data’ has only been touched upon tangentially. And
access to data is becoming increasingly difficult. What can we do if market players close the
access to their data, and, if we find data available, the Research Ethics Board tells us to stop? The
answer is simple: go to the users and ask them for the data. This approach is what the literature
calls data donation. This paper will describe the data donation approach in detail, focusing on
how researchers can access data through users on the current major Western platforms. The
practical feasibility of data donation access will be illustrated based on a domestic pilot study.

The sociological applications of supervised machine learning, already well proven in industrial/
business applications, raise specific questions. The reason for this specificity is that in these applications, the algorithm is tasked with learning complex concepts (e.g. whether a tweet contains hate speech). Supervised learning consists of learning to classify previously annotated (hate
speech/non-hate speech) texts by the algorithm, looking for characteristic text patterns. The
questions that arise are: how to prepare annotation? How can a hermeneutic challenge such as
hate speech recognition be performed by annotators? Are routinely applied, detailed annotation
guidelines helpful? The article also discusses how large companies perform coding on crowdsourcing platforms, and describes AI bias, which in this case means that annotators themselves
introduce bias into the data. I illustrate these issues with our own research experiences.

The web 2.0 phenomenon and social media – without question – not only reshaped our everyday experiences, but they have established an environment for new types of social practices and social actors. The demotization (Turner 2010) effect of such technologies has created entirely new fields where celebrities might emerge from: one of them is videoblogging. Many video bloggers gained great reputation through peculiar micro-celebrity practices (Marwick 2015, Senft 2012), and, as a result, became key figures in distributing ideas, values and knowledge in today’s society. These cognitive patterns are disseminated with a discursive apparatus that is largely based on social media activity, including posts, tweets, self-imagery and the videos themselves, which are tied to a certain logic according to environmental affordances, creating the possibility for fans to interact (share, comment, like, retweet etc.) with artifacts of the celebrity. This mechanism puts the celebrity in a so-called expert system (Giddens 1990) position as they provide adequate schemas of attitude, mentality or behavior. Most importantly, all of these public interactions are accessible for scholars to conduct scientific research. With the help of the SentiOne application this research attempts to reconstruct online networks of video bloggers based on mentions, which either occurred in an artifact (post, video description etc.) or in a fan comment. Apart from the network itself, SentiOne enables us to get insights regarding each individual connection established in it with different types of aggregated data.

One of the world’s largest video-sharing platforms is YouTube, where viewers can comment on
the videos and their topics. The aim of this study is to examine the values and opinions about
unconditional basic income according to the comment sections of several Youtube’s videos which
topic is the previously mentioned UBI which is receiving increasing attention in parallel with
today’s economic and social changes. Our research works with a mixed method, data collection,
storage, sentiment analysis and the bag of words method which were implemented using IT
procedures, while categorization was done through manual coding. The results of the sentiment
analysis show that positive arguments appear to a lesser extent in the comments. Positive
arguments have value characteristics such as inclusion, the principle of the right to exist, justice
and freedom. Among the positive arguments feasibility enjoys the highest support. Negative
category values arise more frequently, so the emphasis on the values of injustice, exclusion,
unaffordability, and performance-orientation is dominant in the analyzed comments.

The research investigates the way users interact with leading topics of the 2018 Hungarian
general election campaign on candidates’ Facebook pages. It expects that the prominent
(immigration, corruption) and campaign-related topics generate more user engagement, while
policy topics and mobilization content are less interacted. It also tests the theory of issue ownership
in relation with user engagement. These expectations are tested on a dataset that includes all
posts (38030 posts) posted by all candidates during the campaign (511 candidates). Topics
are identified by text mining methods. The study demonstrates that corruption, development
policy and campaign are highly engaged topics, while immigration was more interacted only on
opposition politicians’ pages since the followers of pro-government candidates engage less with
immigration-related content. The most surprising result is that a reversed issue ownership effect
can be detected since politicians are generally less successful with their own topics.

My study attempts to explore the lifestyle-based segments of the Hungarian youth through an
innovative methodology based on social media data, incorporating the dimension of digitization
into the creation of lifestyle groups. The examination of the segments’ lifestyle attitudes is
assisted by a review of the related theoretical milieu approaches, international and Hungarian
empirical milieu researches

In the environment of 21st century technology, the transformation of information acquisition
of health care and patients has had an increasing emphasis. Despite the earlier authoritative
doctor-patient relationship, a need for an equal, cooperation-based communication has emerged
and there are so many digital healthcare projects to achieve this (Koskova 2015).
Information acquisition on the internet has allowed patients that based on the increasingly
available medical information they acquire information about their condition, become part of
patient communities, ask for second opinions, and become committed helpers of their doctors in
their disease (Meskó et. al 2017).
This can be especially true for patients with rare diseases, where a diagnosis might take even
a decade, the patient needs lifelong condition maintenance and treatment, if it is available. While the proportion of patients with rare diseases is low compared to the whole of society, the number of such patients is approximately 30 million in Europe (EURORDIS), which means patients
and their relatives need not only a harmonized health care system, but extensive information so
that they can live with the rare disease with less difficulty.
The aim of our study was to present the options of information acquisition in the social
media, focusing on Twitter, via an interdisciplinary and social approach. In this study therefore
we carried out a Big Data based social media analysis based on #Asthma and #CysticFibrosis
databases of the Symplur corporation. This study results contain the complete online communication of 7 years (2012-2019) regarding these hashtags. The analysis has few levels including
semantic research, stakeholder and hashtag review, engagement, and the whole tweet activity
exploration.

The study of the phenomenon of depression is not new in sociology, but since the depression
is becoming a wider social problem, it is still a relevant issue today. In addition to the biomedical and psychological aspects of depression, the sociological perspective is becoming more
noteworthy in the discourse about the causes of depression. In the research of the discourse
on depression, the online texts offer many new possibilities, as the forum’s anonymity and
accessability make the online seeking for help popular. In this research, natural language
processing (logistic regression) was applied to find patterns in the definition of depression
in lay discourses. These methods make it possible to analyze a large amount of text - which
would have been difficult to process with human resources. During the analysis, 67 857 posts of
English-speaking online forums were categorized along the categories of the scientific discourse
about depression. This study presents the first results, which shows logistic regression classifier
performs like the annotators. . Although the research has analyzed English-speaking forums, my
findings may be useful to anyone observing abstract sociological concepts in online texts written
by users.