As a technology journalist, I’ve been fascinated by social media from the very beginning. I loved the fact that the web had evolved into a lively platform for people to engage in conversation and share information with each other; it felt like an exciting movement to be part of. When I hung up my reporter’s hat for a stint as a researcher, I remained just as enthralled, but the focus had shifted. Now I was looking at how computational social scientists were gathering data from social media activity to learn more about human behavior. The web – or at least the open parts of it – became a place to observe homo sapiens at work, at play and, of particular interest to me, at research.
Social media has helped us gain insight into everything from how democracy works to what movies will do best at the box office . By analyzing large datasets of online interactions, social scientists and psychologists can study a population, learn about their habits and even predict future behavior.
During my time as a Masters researcher at the INSIGHT Centre for Data Analytics at the National, University of Ireland, Galway I decided to look at how academics use social media. What I was really interested in was how they shared scholarly articles and other research outputs across social media. Academics have, after all, embraced online technologies as much as the next professional, and I wanted to find out how this was changing the research landscape, from collaborating with others to sharing their research or raising their profiles.
One of the major parts of my research was to look at the emerging field of altmetrics. These metrics typically capture all forms of interactions by any user of the web around research outputs of any kind (e.g. peer-reviewed articles, computer code, datasets etc.), across both mainstream and research-specific social media; including, but not limited to: activity on microblogging site Twitter, blog post mentions or ‘reads’ on online reference manager Mendeley. These alternative metrics have been proposed as a way to capture wider impact of the research lifecycle rather than Journal Impact Factor, h-index or g-index, which are all based on counting citations of peer-reviewed scholarly articles.
Given the plethora of “things” that can be measured, I chose to focus on a particular kind of metric related to scholarly publications in academic journals: these are known as Article-Level Metrics (ALMs), or the tracking of online activity around a scholarly article e.g. a download or Twitter mention.
Luckily, while carrying out my literature review on how academics use social media, I came across Altmetric.com. Founded by Euan Adie, this service is aimed both at publishers and scholars, giving them information on the social media buzz surrounding individual journal articles. I knew this was the best place to begin looking for information: as of May 2013 the Altmetric.com database contained 9.77 million mentions of 1,258,087 journals articles and I’m sure it has grown exponentially since then!
Euan was kind enough to give me access to the Altmetric.com dashboard, where I could explore the kind of activity I was interested in. I could see if a paper was being downloaded on Mendeley, mentioned by online press or in blogs, or how many times it had been ‘liked’ on Facebook or mentioned on Twitter. What really interested me was Altmetric.com’s breakdown of Twitter usage. At a glance it is possible to see what geographic region your audience is coming from, but more importantly, from my perspective, was the demographic breaking users into four distinct ‘types’: members of the public, scientists, practitioners, and science communicators (see screenshot below).
My research centered around how scientists use social media, so it was possible, by using Altmetric.com’s data, to see what kind of papers were being shared by them across Twitter-not just those shared by the general public. This was the foundation for my second study. After an email chat with Euan about this, I was kindly supplied with a dataset of all papers tracked by Altmetric.com. With the help of Excel, I was able to sort them into popularity by Twitter mentions (as opposed to the default Altmetrics score, which can be seen within the famous donut logo).
This combination of ranked journal papers and access to the dashboard meant I could retrieve the Twitter count for scientists and the other demographics.
Suddenly my research began to take shape: I could see which papers scientists particularly enjoyed - and they were different than the ones that the public liked to share!
Suffice it to say that this kind of social media tracking provided by Altmetric.com can yield all kinds of interesting insights that I am confident will help shape the future of science communication. The more we know about how researchers like to use social media, the more we can confidently interpret altmetrics in the context of established citation metrics and tailor social media to their professional needs.