The Exploration and Strategic Services (E&SS) Analytics team assisted in the identification of the key geologic and engineering parameters associated with higher 90 day cumulative oil production for Devon operated horizontal wells in the Woodford Shale in Oklahoma. The project deliverables were a list of effective parameters, their effective ranges, and optimal combinations. Achieving this goal required the collaboration of geoscientists, engineers, technicians, and analytics professionals. Innovative solutions were developed for data capture, data integration, and data visualization. The E&SS Analytics team used Statistical Analysis Software (SAS) Enterprise Guide for data integration and preliminary analysis and SAS Enterprise Miner for the creation of predictive models. In addition to the SAS software, we also implemented a modified version of SAS data analysis workflow, Q-S.E.M.M.A.-P. The newly integrated data was distributed to the team via a Spotfire project. Predictive models were developed as an addendum to the original scope of the project. The data analysis techniques for this projects included univariate, bivariate, and multivariate analysis, outlier detection, Spearman correlation analysis, multi-collinearity detection and resolution, and statistical t-Tests. The predictive models under evaluation included decision trees, gradient boosting decision trees, neural networks, and stepwise multiple logistic regression.

The effective collaboration of petro-professionals and analytics professionals not only produced the deliverables for the project, it also prototyped the automation of data integration of geoscience and engineering data, the deployment of predictive models, and the implementation of the Q-S.E.M.M.A.-P analytics workflow.

This content is only available via PDF.
You can access this article if you purchase or spend a download.