More on text embedding driven sentiment analysis

Another sentiment visualisation using just text embeddings, tracking changes over time. In this case a subset of North Sea Transition Authority offshore license reports between 2008-2017 for the word vector ‘seal’. These types of techniques can support analogues and insights for renewables, carbon capture and storage sites, subsurface radioactive storage, oil & gas exploration, mineral exploration, geohazards etc.

Arguably these data driven techniques support a “Khunian like” philosophy towards discovery. Through identifying anomalies (in the sense of data that does not fit with our existing individual or collective understanding) enables us to question the validity of previous assumptions and generate new research, business questions and theories. Big data==small patterns!

Of course some trends can be data artefacts and correlation is not causation; but the more we look, the more we are likely to see.

#subsurface #oilandgas #carboncapture #naturallanguageprocessing #artificialintelligence #analytics #geosciences #geology