In a previous post, we have explained how we collected a list of about 1000 articles related to climate change, and presented several measures of the associated discussions. Here we want to study the network of hyperlinks between these articles.
More specifically, we build a graph in which each article is a node, and a directed edge connects node A to node B if the corresponding articles are linked, i.e. if article A contains a link to article B.
As in Wikipedia each article represents an encyclopedic entry, and thus an entity, each link can be interpreted as evidence of a relationship between the entities which are the subjects of the articles. Following this idea, for example, the network of hyperlinks betweeen biographies in Wikipedia can be interpreted as a network between notable persons and historical characters, as shown in Aragón et al. (2012).
In the present case, each article represents an issue, an event or an actor related to climate change, and the network can reveal how these entities are related to each other according to the Wikipedia community.
Network of hyperlinks between Wikipedia articles related to climate change
In this figure, drawn with Gephi, the size of each node depends on its pagerank. Colours have been assigned to nodes and arcs according to a clustering algorithm (modularity maximization), so it is possible to get an idea of the main groups (clusters), in which articles are strongly interconnected. To better visualize the network, you can download the pdf version (searchable).
On the bottom-left of the figure we observe a purple cluster which includes articles Global warming and Climate change and is mostly focused on the scientific debate on climate change, and the associated controversies and actors (e.g: Global warming controversy, An inconvenient truth, List of scientists opposing the mainstream scientific assessment of global warming, Al Gore). This cluster has no clear boundaries and is strongly interlinked to a turquoise one on the bottom, focused on global models for climate, and centered around the Intergovernmental Panel on Climate Change.
If we keep moving towards the right, we can see a cyan cluster grouping articles associated with climatology and the effects of global warming, such as Ice Age, See level rise or Greenhouse effect, and a dark blue one mostly related to Climate change mitigation and Geoengineering.
On the top we can distinguish two clusters related to energy and climate change: a red cluster on the right, centered on Greenhouse gas, Fossil fuel and Renewable energy, and a green cluster on the left, associated to the Kyoto protocol and policies for emission reduction.
Close to this we find the last cluster (mustard, on the left), grouping articles related to international treaties and conferences on climate change (the most relevant one appears to be the United Nations Framework Convention on Climate Change) and more in general Individual and political action on climate change, including some relevant actors such as the World Bank and Greenpeace.
Beyond showing how articles can be grouped into different clusters, the figure also reveals which articles are more relevant in this network, according to the pagerank. In order to investigate further which are the most relevant nodes in the network, we will apply and compare several centrality metrics in the next post.
Aragón, P., Kaltenbrunner A., Laniado D., and Volkovich Y. (2011).
Biographical Social Networks on Wikipedia – A cross-cultural study of links that made history,
WikiSym ’12 – 8th International Symposium on Wikis and Open Collaboration.