A Big Data Approach to Energy Transition: Exploring Pathways in Textual Corpora

The concept of ”energy transition” refers to a fundamental socio-technical transformation towards more sustainable ways to produce and utilize energy. Energy policy research building on this concept argues that there is no through one singular “right” way of energy transition. Instead, it is assumed that there are various transition pathways occurring in different contexts and with emphases on the different dimensions associated with decarbonisation. The benefit of studying energy transition through the lens of transformation pathways is that it allows to address the uncertainty and complexity of sustainability challenges more dynamically. Geels and Schot (2007) proposed a typology of four transition pathways: transformation, reconfiguration, technological substitution, and de-alignment and re-alignment. Foxon (2013) set out three core transition pathways: Market Rules, Central Co-ordination and Thousand Flowers. Most recently, Rosenbloom (2017), defined three core conceptions to illustrate pathways for sustainability transitions: (1) biophysical, (2) techno-economic, and (3) socio-technical. As such, pathways emerge as a useful analytical concept with which one can engage with the plurality of low-carbon possibilities. While these conceptual developments provide entry points for empirical research, they are not necessarily exhausting concepts or accurate descriptions of policy, media, business or other expert communities discussions that shape actual policy-making. Moreover, the existing typologies were built on examples stemming from certain contexts (mainly, liberal democracies with diversified economies), and it remains unclear how well they perform in different political and economic conditions. This paper proposes a methodology for exploring transition pathways based on machine learning techniques. We demonstrate how the proliferation of textual big data coupled with development of computational textual analysis methods gives us a unique opportunity to study the concepts of transition pathways in a new and exploratory way. We illustrate our method on large corpora of policy documents, media and expert discussion from the EU and Russia.