What Wikipedians Can Learn from Tracking Hashtags on Social Media

Wikimedia platforms are not isolated entities, nor are the editors, readers, contributors and staff around the globe. As a community, we are connected through our passion for accessible technology and free knowledge. However, societies are evolving with and within other digital technologies, and this tendency should concern us. Not necessarily because we are “left behind” in a technological landscape where Artificial Intelligence and social platforms are taking over, but because the community’s future depends on how well we understand the momentum and pass our mission to new generations.   

As an intern in the social appropriation of knowledge team at Wikimedia Colombia, I began research on how to spot “in the wild” manifestations of denialism or disinformation on social media. Our purpose was to identify the strategies, narratives, and nuances of denialism, primarily in Latin American contexts. The more we learned about what disinformation is and how it is connected to other digital phenomena, we started questioning how to contribute in a productive yet innovative way to Wikimedia platforms. Our proposal was to use open-sourced tools to capture and visualize online data produced on social media and then use this material as a source of information—and motivation!—for Wikimedia’s efforts to improve knowledge integrity. We found this approach especially useful for the mission of organizing and exchanging free, trusted knowledge beyond Wikimedia

Our final research idea was tracking how climate disinformation manifests on social media, using the discussions on X around fracking as a case study. The results of this process was materialized in several pieces (slides, a research article, and data visualizations) which are available on Commons if you want to skim them through (in Spanish).

We learned a lot along the way, and here are five lessons (we wished we knew earlier!):

1.    You will find useful info where you least expected it

When looking for disinformation online, where do you seek? On chatgroups, fact-checkers websites, Facebook groups? We know there is all types of disinformation flowing around, but how do we catch them? Unless we are familiar or immersed in digital spaces where the main content is disinformation, tackling this becomes a problem. Take fake news as an example. If we are searching for potentially fake news that might be present in Wikipedia’s articles, it would be naïve to search for the term ‘fake news’. Fake news do not label themselves as such. So, we might need another strategy. 

We came out with one approach that worked for us. When looking for online phenomena such as disinformation, instead of focusing on the fruit, go for the rind. What we mean is that it is more productive to look out for orbiting issues that are somehow connected with our topic of interest than addressing its specificities. Imagine it as throwing a big net on the river instead of using a fishing rod. 

Our initial perception about climate change disinformation in Colombia and Latin America was very limited; we started working on controversial cases around the topic but the data we gathered each time was kind of vague and inconclusive. When we broadened our perspective, we found the right track. By thinking our topic in terms of related discussions around climate change as a big phenomenon, not necessarily as disinformation cases, we found a discussion about dependence on hydrocarbons. So, we decided to track the fracking issue, and indeed we identified denialism and climate change disinformation. 

Even if this strategy might need to be reformulated in each case, we advise you to throw the net if you haven’t caught what you were looking for. It may surprise you! 

2.    Scale your research

The more we follow the specialization culture, which pressures us to segment knowledge and techniques, the more we overlook the relevance of our own work in other areas. We have internalized to keep within bounds and narrow our findings. To focus exclusively on one problem. However, there are times in which our data allow us to go against the grain. 

For our research, we tracked #fracking on X and recollected 3.804 tweets that included in its content at least one time that hashtag. While analysing our data we found networked publics that participate at different intensities and frequencies. Alongside the climate change denialists, there were other publics that addressed the fracking discussion around policy flexibilization, natural disasters, authority’s accountability, tensions during campaign periods, and so on.  

Considering the connection between these publics as part of our corpus allowed us to understand the relevance of our research with other topics of interest. We found what we needed, but our findings were not created in a vacuum, they also expressed fracking as part of a public agenda dependent on sociopolitical contexts and regulations. More interestingly, the peak of this general discussion was during presidential debates which also were the times of most propagation of disinformation around climate change. Take for example the presidential debate between Donald Trump and Kamala Harris in September 2024. Related to each candidate’s posture around oil and gas production, two hashtags —#DrillBabyDrill and #NoVotes4Kamala—included political commentaries but also misleading information. 

Following previous research findings, we expected to find some connection between climate change denialism with conservative American politics. However, we did not expect that when the fracking issue is mentioned, users who claimed being victims of the consequences of fracking in their lands used the same hashtags (#fracking, #NoVotes4Kamala) than the denialists and supporters of fracking. We couldn’t have arrived at this finding if we had isolated the cluster of denialists from the other publics and focused exclusively on disinformation content. 

We did not end here. For the upcoming elections in Colombia in 2026, we will track social media during live debates and corroborate whether the same integration of partisan/programmatic hashtags and climate change disinformation occurs. 

 3.    Dare to change the recipe!

You must be questioning how the findings of this research relate to Wikimedia’s platforms or endeavours? First, valuable initiatives within the community, such as the Anti-Disinformation Repository, bring attention to the relevance and urgency of fighting disinformation online and the role of Wikimedia as an antidote. Our research on climate change disinformation is not a new topic, but it does apply a new approach borrowed from critical media studies, which is based on the premise of follow the natives, meaning understanding first the medium and how the phenomena occur within it and then put your insights at the service of what you want to tackle.

What we propose is that we need to be familiar with how disinformation flows on the Internet, not only on our platforms, to create effective tools and strategies to fight it from our corner. This is a crucial step for upgrading our approach to disinformation as a community because it will strengthen our research skills, but also our role as active digital citizens. Otherwise, we face the risk of investing valuable time, ideas and resources in less impactful endeavours. 

Second, during our research we also inquired ourselves about our contribution to the platforms. How can we enrich disinformation related articles using what we have learned? We discovered that in several Wikipedia articles there is little to non-information around strategies of disinformation online— a well-documented and popular field on media studies. Instead, the articles focused on the content, for example, whether there are sources to dismantle a statement, or giving context around a manipulated figure. We believe that both approaches —strategies and content— might be presented together on our platforms. This will provide users and communities with both fact-checking insights and media literacy skills. 

We are working on it!