in the online journal called Somatosphere. And the piece has the same title as my presentation here, how we tweet about Coronarios and why a computational anthropological mapping of political attention on Danish Twitter during the COVID-19 pandemic.

I should say that out of this group of an interdisciplinary group, I have a background in science and technology studies myself. So I was elected the proper representative, so to speak, in this context. You will notice, of course, that unlike I think everything we've heard so far, the topical focus here is not on science,

but the methods and I think the broader scope of my presentation certainly pertains to STS-related issues. It's important for me to say that we claim little

originality in the main idea to use social media data, in this case Twitter data, to try to understand public issue attention dynamics during the COVID-19 as a focusing event, as we call it. This has been done in a lot of contexts in other settings.

I'm sure it's been done also in other geographical settings than our Danish setting. As always, and since this is an STS meeting, I thought it would be appropriate to point to some of the practicalities locally in our research setting that made this response

that we call a computational anthropological response feasible in the first place, meaning, first of all, the fact that we happen to be an interdisciplinary research group, adding data scientists through to anthropologists with the capacities to do large-scale digital methods work

in the sense that this has been discussed in the last let's say 10 years, also in the STS literature in terms of digital methods. And also, as others of you have mentioned, the ambition, if you like, on our part to perform what I call here

real-time research, attempting to do digital public mapping and make this mapping available, not in real time, but relatively quickly through various kinds of public outreach, including the GitHub that you see here

and that I, of course, hope you will have a chance to visit at some point, although I do apologize in advance that a lot of it is going to be in Danish. Of course, there's somewhat of a point to be made there in terms of the kind of public that we are trying to engage with our research.

This is a first take on trying to landscape the Danish COVID-19 Twitter space. What you see here is a co-hashtag network based on approximately 140,000 Danish language tweets

cleaned up, so to speak, according to language and according to core search terms in an iterative process. It's important to note, of course, the contextual factor that Danish Twitter is used by approximately 10%

of the population according to recent surveys, predominantly by what we could call a high status, politically savvy and oriented community. So of course, we don't take this as representative of public conversations at large.

Having said that, what stands out quite clearly are three thematic or issue clusters that I think resonate, in fact, also with some of what Rodrigo just presented. There's a health policy cluster in the middle in blue, apologizing again that this is in Danish, so you won't be able to make it out.

These are Danish language hashtags. There's a health policy related cluster in blue. There's an economic policy cluster in purple. And on the right, towards the top, there's what we call a civic morality cluster in orange, meaning basically civic morality oriented reactions

to the government enforced and implemented lockdown of public institutions that came into place in Denmark in mid-March. And I forgot to mention that what you're looking at here is the total period from the end of February

to end of April, so in other words, covering that lockdown period in particular. We did look at dynamic networks and these networks over time. This is another way of showing some of those temporal developments.

This is, again, I think resonating with some of what Rodrigo just presented in his talk. One thing to notice here, I think, in the upper graph is, again, if you focus on the orange line, this is what we call the civic morality reactions.

So these will be tweets like injunctions to stay home, to wash your hands, to not hoard consumer goods and things like these. And reactions like this pick up rather dramatically

after the government imposed lockdown and at some point towards the end of March, in fact, constitutes the most prevalent conversation, if you like, or topic being discussed on Danish Twitter. So again, pointing to, I think, some of the moral

and, as we would say, also affective dimensions of how this techno-scientific and social situation is being handled in the Danish public. This is being unfolded a little bit more in the bottom graph, which I won't go into much. I suppose one fun fact, if you like,

to notice is just the spike right after the lockdown that you can see in the red bar. And that spike indeed is injunctions to other fellow citizens not to go hoarding

for consumer goods following the government imposed lockdown. So again, a clear sign of civic morality, as we would call it being expressed on Twitter. This kind of analysis led us to focus more, not simply on what was being said on Twitter,

but also how it was being said. And again, I think there's a link to what Rodrigo just presented. This is not a sentiment analysis, although it resembles it in some respects. We coined the term affect analysis for what you see here. It's basically similar in the sense

of using dictionary-based prediction models that we have cleaned up and categorized tweets accordingly with a basic typology coming out of standard psychology in terms of emotional registers present

in the speech acts of the tweets that we analyze. Again, the clear pattern that stands out, I think is clear here. We see conversations pertaining to and being expressed in a register of trust or what we call trustiness being very prevalent

after the announcement of the lockdown. If we dig deeper into this, and again, this will be standard knowledge to those of you doing these kinds of methods, you will see that trustiness includes indeed expressions of mistrust.

In other words, this should not be interpreted, we argue, in terms of sentiments or emotions actually unnecessarily helped by the people doing these tweeting acts, but rather should be read as an affective expression. We argue of a discursive norm

of proper performances of citizenship under the COVID-19 situation, which in a Danish, or we could perhaps say Nordic, Scandinavian welfare state context seems to attain the tone or the modality

very much of trust as a public matter of concern that a lot of public actors relate to. And of course, they relate that partly in the sense of trust in science, but they also relate that in the sense of trust in government. And then indeed, trust in each other as citizens to perform according to the scientifically based injunctions.

And in a forthcoming piece that's coming out in current anthropology, we argue that this might in a further sense be interpreted as a form of Danish bio-political nationalism. This is a term that we pick up from other conversations

actually based in East Asia on public reactions to the COVID-19 pandemic. Basically again, in the sense of this being indicative of a certain affective mood that citizens in this Twitter space are expected to perform according to,

in this case being trust. And of course, there's a broad literature on the importance of trust in a Nordic Scandinavian welfare state context like the Danish that we discussed this vis-a-vis. May I interrupt you? Can you please finalize your talk?

I'm going to end right now. Thank you for letting me know. I want to end here just by raising a few questions or observations basically on what to make of work like this that happens as I call it here in the middle of Corona concerns

and in the COVID-19 meantime, in terms of broader, let's say STS related implications. I see two topics that I would also like to discuss with you, one has to do with what I'm calling here the politics of digital methods. Obviously, all sorts of methods have been

and are being deployed to know and make claims about public reactions to the COVID-19 pandemic. Our digital methods sits vis-a-vis more, let's say strictly computational approaches for which I would include sentiment analysis for instance, but of course also a hinterlander survey approaches

and so on. And I think there's an interesting question as to how to take stock of these different methods for knowing public reactions. And then secondly, I certainly for one has become aware of the relative scarcity of strictly speaking STS informed approaches to digital media publics writ large

as a topic of sociocultural studies. We draw notions like affective publics that come out of anthropology and sociology and so on. But I think there's an interesting questions as to what should be the particular STS contribution to theorizing the kinds of public reactions on digital media that we see here.

Thank you.