Blueback Investigator

Blueback Investigator is a cross-domain data analysis plug-in for the Petrel* E&P software platform.

Data analysis in Petrel is often challenging for users; Blueback Investigator provides a rich data analysis experience and enables more workflows for the user.

Statistical analysis tools can be applied to understand well and point data, 2D and 3D seismic, 3D model and simulation properties and surface attributes.

Some example workflows using Blueback Investigator include:

• Identification of incorrectly loaded data
• Comparing trends between wells
• Understanding data distribution and trends prior to building a reservoir model
• Predicting missing data
• Examining seismic attributes in different spatial regions within a volume

Investigate data in five dedicated, unique, feature-rich windows; histogram, crossplot, matrix, parallel coordinates, and 3D plot window.

Investigations can simultaneously be shown in multiple windows, each with a different view, allowing you to identify and investigate data relationships. Descriptive statistics can be displayed, either as a table or visually, in all windows and updated in real time as visualization options are changed.

Additional insights can be gained through the spatial context, which enables visualization in native Petrel windows alongside native Petrel objects adding invaluable context.

Windows can visualize investigations containing:

• Log data from 1000s of wells in an unconventional field
• 10s of seismic attributes from the same volume
• Model properties that vary across multiple simulation dates
• Well log/upscaled/model property/simgrid QC

Application of data science techniques is growing rapidly within the Geoscience domain.

Integration with Python™ through Cegal Prizm allows the Blueback Investigations to be accessed in real time from a Python environment. This allows you to augment existing functionality and well-known techniques with the latest state-of-the-art algorithms and models to delve deeper into the data. Application of the Python regression or classification models in Petrel requires no knowledge of Python – all Petrel users can easily use the model directly within the Blueback Investigator plugin.

Creation of the Blueback Investigation forces some "data cleaning" by ensuring that all properties from different datasets have consistent units, spatial location, and sampling, as this should have been considered while loading the data into Petrel. Data from different domain objects is combined based on Petrel templates.

Blueback Investigations can be plotted directly from Python using a simple API, significantly reducing the amount of code required to display complex visualizations.

Collaboration between the Python user and geoscientists is improved as these plots are exactly like the windows in Petrel, giving confidence to geoscientists and aiding discussions. The appearance of the data will be consistent across all related visualizations.