Tapping into data from Bluesky

Altmetric has expanded its tracking capabilities by integrating data from Bluesky, the rapidly growing social media platform with over 24 million users, as a new attention source. We spoke to Miguel Garcia, Product Director, Altmetric, to get insights into why the decision is an important one at this stage.

Why was Bluesky chosen as a new Altmetric attention source?

We already collect social media mentions to research outputs from a number of sources. X (formerly Twitter) is probably the most relevant one, but we collect data from other social media data sources as well. However, we are noticing that there is a movement away from the status quo. There is a changing environment with a more diverse pool of platforms for people to discuss research on social media. And although these have existed for a while, they’re just becoming a bit more mainstream. 

We found that  Bluesky in particular had an amazing potential.  It is very similar to X in some ways. So we expected people to be familiar with the way of engaging with that platform. It sounded like a natural place for people to move to when continuing research conversations online. 

So from a user experience perspective, and from a user interface perspective on the platform, we thought Bluesky would be a place that academics would go and our research so far shows we were right. Moreover, it was also a platform that was ready to give us that data and was easy to integrate with. They have an infrastructure that is designed to be open, so it is ready for interoperability unlike many of the other social media tools that tend to be very closed, commercially focused and not as easy to integrate. 

Moreover, at Altmetric our goal is to track attention and impact to research outputs across platforms and we do not do this exclusively for social media. We are constantly reviewing our data sources and adding more that complement the ones we have.

For example, the Clinical Guidelines that we recently released or Wikipedia, where we continuously look to add new languages. 

What gap does Bluesky fill compared to existing data sources from social media platforms?

Many of the data sources we currently have in Altmetric Explorer reflect the historical nature of social conversations. We have some social media like Google+ for example that no longer exists. But we still retain the historical data that we have captured in the past, this is the same across X, Facebook, Reddit, etc. We think that given the changing environment, and the change social media is going through right now, the gap between historical and current data would start to widen. So, while it will be impossible for us to capture every single research-related conversation that exists, we want to get as close as possible. Right now we need a more diverse pool of social media sources to have a better chance at getting the complete picture of what is being said about research output. Despite the fact that there is a very significant amount of research output mentions on X, we believe that gaps will appear in the future. In fact, we’ve noticed that indeed the number of X mentions are slightly going down. So we made the decision to diversify our pool of social media, and especially to bet on Bluesky. 

One thing that we saw now as an opportunity while ingesting BlueSky was alternative text (alt text). So when you post a picture in Bluesky you can add alt text to be displayed on that picture and this is mainly for accessibility. And what we’ve realized is that some users of Bluesky added the URL of the publication as alt text of a picture. That’s something we were not doing before and we thought ‘here’s a good idea: let’s track mentions from alt text as well’. So as we are indexing more data sources, we also see new ways that people are engaging with social media. 

Will Bluesky data be treated differently in the Altmetric Attention Score? Will its contributions be weighted the same as X, or will it have distinct metrics?

So, a post on Bluesky will count as much as an X post, and a repost and a quote post should count similarly to that of other social media channels. And we should also have the same conditions for Bluesky.

So, let me clarify: A post is what it says, a repost is when you repost something that someone else has posted without any further content, a quote post is when you do that but you add your own text. So we consider those things to be slightly different. One thing is for you to click a button and repost something. Another thing is for you to share something and add your own words on top of it. That is contributing to the discussion. It’s not just amplifying the content that has already been posted.  So we count those as different things, but also we have certain limitations around social media posts in the way that we count them towards the augmented attention score and this is in order to prevent gaming. This is vaguely explained in our documentation. It’s not explained in granular detail because there are rules that we put in place in order to prevent gaming. So if they were too transparent then people would be in a better position to game these metrics and that’s what we’re trying to avoid. 

All of this is accounted for as conditions of how much these mentions count towards the Altmetric Attention Score. So, the same limitations will be placed in Bluesky precisely to control the same kind of behaviors. We continuously monitor and evolve our practices around fighting bot behavior, ensuring best practices and good faith around social media sharing and how much it contributes to the Altmetric Attention Score. So from that perspective, Bluesky presents yet another opportunity for us to see how users behave when discussing academic research on social media.

Does Bluesky provide new ways to demonstrate impact or reach untapped audiences?

From the perspective of a data provider that is collecting data from social media, Bluesky presents the same value as any other social media channel. It presents an opportunity for users to engage with and discuss research. I am going to abstain from commenting on the recent developments of X and how institutions feel affected by it.

It’s clear that many users have left the network, not only the institutions but the researchers themselves; we have recently run a survey for Altmetric Explorer users and we’ve asked them which social media network they use most frequently to share, or engage with scholarly research, and still a good 58% of them said that they were using X. But 42% said they were using Bluesky. We should be aware that a user might have more than one response.  So a sizable amount of users seem to be using Bluesky and LinkedIn to share research. Now there was another question that asked if your usage of X changed in the past 10 months and 61% said that it has decreased, 14% said they’re not using it, this means that 75% of our users that responded to this survey are either not using X or see their usage decreasing.

So clearly there are opportunities in another social network that X is not giving them. A vast majority of respondents said they were concerned about trust and platform integrity, and over 50% also said they were led by the community migrating to other platforms. What we see here is that there is group behavior in place. So if an entire community starts using another venue to share and engage, then the people participating in that community feel forced to move as well. However, the main reason was concerns about trust and platform integrity.

How does the inclusion of Bluesky position Altmetric in comparison to competitor solutions?

To the best of my knowledge, we are the first product to integrate with Bluesky in this way; some of our competitors have even stopped indexing X.  So we are definitely the most complete source for altmetrics data regarding social media and Bluesky is only going to enhance that by allowing us to be pioneers in listening to these researchers’ voices and what they’re saying about research on social media.

It reinforces our position as a unique solution to properly assess the interest and impact of research beyond citations alone, which we’re also getting from Dimensions, and allows a truly holistic assessment of research beyond the classic citation analysis.

Do you see Bluesky as a potential replacement or complement to Twitter in the long term?

Right now, according to our survey, we’ve seen part of the researcher community is migrating to Bluesky and so it was necessary for us to capture this attention from Bluesky as well. However, we still see a number of discussions about research happening on X. So from purely a data perspective, there would be no reason for us to stop indexing X.

If in the future we do see an erosion of these discussions in X, we would consider dropping it as we do for any other source. We constantly review the sources that we are ingesting. We constantly look for new sources and we look at the ones that we’re currently indexing to think if it’s still worth it to continue capturing their input. But so far we do not have any plans to deprecate our ingestion of data from X because we still see a lot of those discussions happening there.  But the future will tell if the strategy should change and I should not close any doors there.

It’s been a busy year for Altmetric, what’s next?

I am very proud of the work that the Altmetric team has been doing in this field. This is the second data source added to Altmetric this year. But it’s not just about adding data raw, it’s about showing insights from this data.

One thing we have done earlier this year is to present a timeline with the mentions that go to a paper broken down by source since the paper was published, and that helps you compare the type of mentions and the number of mentions that come into that paper per source.  So you’ll be able to then compare how many mentions come from X versus Bluesky versus policies or clinical guidelines or news. And I think adding more data complements those broader analytic features that Altmetric has.Besides that, we aim to continue to develop additional insights on these sources and social media in particular. So our next investment is the sentiment analysis of these discussions, knowing not just that this research has been mentioned or what exactly was said, because those are things that you can currently do in Altmetric Explorer, but having a sense of the sentiment around those comments, and those posts. That discussion can be particularly useful when you have a large number of them. When you are just looking at a paper with two mentions, it is very easy to click on them and see what has been said and who said it. But when you’re analyzing papers that have had a massive discussion online, it gets really hard to go through dozens, sometimes hundreds or thousands of pages to look at every single comment. it’s just not humanely possible. So we aim to provide an aggregated analysis on this. Our next step is to apply the sentiment of these mentions to Bluesky and X for analysis.

If you want to learn more about Altmetric, don’t hesitate to contact the Altmetric team.