Energy Data Science Meetup

Share on twitter
Share on linkedin

Challenge #1 Mitigating Land & Lease naming issues across or within a data set. This is a common problem for most dealing with oil and gas data. In this approach we apply a bag-of-words model to quickly determine the range of uncertainty regarding naming consistencies. The approach leverages a TERR (TIBCO Enterprise Runtime for R) & Spotfire. The machine learning technique (text analytics) can augment engineering workflows in a variety of ways. In our example we reveal how this technique allowed for upstream data to connect to midstream data providing a more completed picture of Gas Yield/Shrinkage Accounting for a Gas Operator. Challenge #2 In mature and highly developed (read crowded) fields better understanding of the areal deposition of wells can provide key insights for infill drilling, sweep efficiency, interference, and more. In this approach we apply voronoi tessellation to provide both visual insight into field/well behavior over time as well as stage the data for further data science applications (such as kriging). The approach leverages an R Cran Package (Open R), TIBCO Statistical Services Server (TSSS), & Spotfire. This geospatial technique can augment engineering workflows in a variety of ways. In our example we reveal how this technique may be applied to the Eagleford in Texas where this geospatial data science technique is the starting point for predicting well performance based on rock properties (core data). About is a data science boutique located in Houston. They focus on building advanced analytics packages for oil and gas assets across the globe. Clients include both super majors and small independents. They are a TIBCO Spotfire partner and published SPE authors on the platform. About the Presenter Troy Ruths, Ph.D. Chief Data Scientist Troy Ruths received his BS in Computer Science from Washington University in St. Louis and a PhD in Computer Science, with a specialization in Computational Biology, from Rice University.]]>