Realize the potential of your unstructured data with KNOWLEDGE EXTRACTION

The data digestion provides easy access to information in a fast and efficient fashion; it is orders of magnitude faster than current manual approaches.

The digitization process makes it possible obtain new knowledge and in-sights that is very difficult or even impossible to achieve using unstructured data in its existing format.

The knowledge extraction is done through many types of higher order analysis. Examples are:

• Drilling history: A knowledge graph is used to visualize dependencies between wells and understand the drilling history in an area. The analysis is constructed by interrogating each well in the database and investigating the degree of correlation with the remaining wells

• Rock physics modelling: As all tables in well reports are available as .csv files the parameter selection building a rock physics model can be done with higher confidence. Clay volume cut-offs, end member definitions and depth trends are more easily picked and validated. It allows for a geological constrained rock physics models that are supported by huge amount of data, as opposed to a limited block-by-block approach, where data is sparse

• Sedimentology: Using t-Distribute Stochastic Neighbor Embedding (t-SNE), which are dimension reduction techniques suited for visualizing high-dimensional data, thin sections from several hundred well reports can be clustered without any prior knowledge with similar features regardless of their geospatial proximity. As any identified images are tagged the origin is fully traceable back to the specific page in the original report