Abstract
In the INVENT project, we aim to develop new perspectives and methodologies for capturing the wider societal value of culture. To understand how Europeans view culture and its societal values, we follow a bottom-up approach to researching and mapping cultural diversity in Europe. In doing so, we employ
– among other things – data scraping techniques to collect digital content and gain a better understanding of how Europeans express themselves online about culture. Two main research questions guided our investigation in the first data scraping phase: First, in mapping the cultural ‘Twittersphere’ in all INVENT partner countries, we asked which topics are associated with the keyword
‘culture’ in conversations on Twitter during the years 2019 (pre-Covid19) and 2020 (during Covid19).
Second, we were interested in the types of actors that engage in conversations about culture on Twitter and the extent to which these interact with each other.
The data scraping process began with the social media platform Twitter in January 2021 using R-programming language. We decided to focus on Twitter for several reasons. First, Twitter is prominently used for sharing news stories, but also enables the emergence of stories from below, based on individuals’ experiences, opinions, and struggles, making it a highly relevant source for
studying understandings of culture from the bottom-up. Second, Twitter offers one of the most accessible websites for research in the social sciences. In 2021, the platform introduced an advanced Academic Research product track that facilitates access to the platform’s historical archive for qualified
researchers. To scrape relevant tweets, we used the keyword “culture” in the INVENT project’s respective country languages; in some cases, the sample was further extended with English-language “culture” tweets as well. We focused on tweets posted from users’ home locations in each country and analyzed the tweets through topic modelling, a powerful and effective method for detecting
relevant hidden structures in the data.
We identified eight thematic topic groups across the nine INVENT countries: Tweets and conversations about “Film/TV/Theatre”, “Literature/Arts”, “Music/Concerts/Festivals”, “Policies and Politics”, “Identities and Boundaries”, “Society and Inequalities”, “Economics/Business/Work”, and “Spaces and
Places”. As these topics are identified in almost all countries, they are at the heart of discussions related to culture in Europe. Additionally, we detected several country-specific topics that are unique to the particular context in each country, such as conversations on “MeToo” or “Health”. Such context-
specific cultural topics could be of interest to cultural policymakers, but they also highlight the need to acknowledge cultural discussions on the regional and national level. Our analysis shows that there is a widely shared basis of culture that refers not only to the arts, leisure, and creativity (“culture as arts”),
but also to more elementry aspects of human life such as politics, inequality, and economics (“culture as way of life”). Another similarity between many countries, albeit to varying degrees, is the relationship between culture and the emerging Covid19 pandemic, where Twitter users discussed the
shutting down of cultural venues and the policies implemented to save at-risk events and venues. The most active users in our data are media and cultural organizations and individuals such as journalists and artists, that account for a significant share of tweets in each country. This means that the
perceptions and viewpoints of specific organizations and professions are to some extent overrepresented in our data. Still, we found a large variety of users in each country who contribute to cultural discussions on Twitter, indicating that analyzing Twitter data is indeed useful in studying culture in a bottom-up fashion. Despite choosing a general keyword, our approach restricted the data
collection to tweets that explicitly mention the word “culture”, and posts about other cultural offerings were, therefore, not included in our data. This report is the first phase out of three in our web scraping
research. Based on our findings, for the next phase, we are considering looking into active users or particular cultural offerings that were not reflected in our current findings.
– among other things – data scraping techniques to collect digital content and gain a better understanding of how Europeans express themselves online about culture. Two main research questions guided our investigation in the first data scraping phase: First, in mapping the cultural ‘Twittersphere’ in all INVENT partner countries, we asked which topics are associated with the keyword
‘culture’ in conversations on Twitter during the years 2019 (pre-Covid19) and 2020 (during Covid19).
Second, we were interested in the types of actors that engage in conversations about culture on Twitter and the extent to which these interact with each other.
The data scraping process began with the social media platform Twitter in January 2021 using R-programming language. We decided to focus on Twitter for several reasons. First, Twitter is prominently used for sharing news stories, but also enables the emergence of stories from below, based on individuals’ experiences, opinions, and struggles, making it a highly relevant source for
studying understandings of culture from the bottom-up. Second, Twitter offers one of the most accessible websites for research in the social sciences. In 2021, the platform introduced an advanced Academic Research product track that facilitates access to the platform’s historical archive for qualified
researchers. To scrape relevant tweets, we used the keyword “culture” in the INVENT project’s respective country languages; in some cases, the sample was further extended with English-language “culture” tweets as well. We focused on tweets posted from users’ home locations in each country and analyzed the tweets through topic modelling, a powerful and effective method for detecting
relevant hidden structures in the data.
We identified eight thematic topic groups across the nine INVENT countries: Tweets and conversations about “Film/TV/Theatre”, “Literature/Arts”, “Music/Concerts/Festivals”, “Policies and Politics”, “Identities and Boundaries”, “Society and Inequalities”, “Economics/Business/Work”, and “Spaces and
Places”. As these topics are identified in almost all countries, they are at the heart of discussions related to culture in Europe. Additionally, we detected several country-specific topics that are unique to the particular context in each country, such as conversations on “MeToo” or “Health”. Such context-
specific cultural topics could be of interest to cultural policymakers, but they also highlight the need to acknowledge cultural discussions on the regional and national level. Our analysis shows that there is a widely shared basis of culture that refers not only to the arts, leisure, and creativity (“culture as arts”),
but also to more elementry aspects of human life such as politics, inequality, and economics (“culture as way of life”). Another similarity between many countries, albeit to varying degrees, is the relationship between culture and the emerging Covid19 pandemic, where Twitter users discussed the
shutting down of cultural venues and the policies implemented to save at-risk events and venues. The most active users in our data are media and cultural organizations and individuals such as journalists and artists, that account for a significant share of tweets in each country. This means that the
perceptions and viewpoints of specific organizations and professions are to some extent overrepresented in our data. Still, we found a large variety of users in each country who contribute to cultural discussions on Twitter, indicating that analyzing Twitter data is indeed useful in studying culture in a bottom-up fashion. Despite choosing a general keyword, our approach restricted the data
collection to tweets that explicitly mention the word “culture”, and posts about other cultural offerings were, therefore, not included in our data. This report is the first phase out of three in our web scraping
research. Based on our findings, for the next phase, we are considering looking into active users or particular cultural offerings that were not reflected in our current findings.
| Original language | English |
|---|---|
| Publisher | Invent consortium (European Inventory of Societal Values of Culture) |
| Number of pages | 143 |
| Publication status | Published - 31 Jul 2021 |
| Publication type | D4 Published development or research report or study |
