Guest Post by Ky Kiefer
We often design hydraulic fracturing treatments around historical success. We look to the successes (and failures) of “wildcatters” of the past and attempt to replicate treatment designs. Once we have a baseline treatment that tends to perform within a formation, we strive for continuous improvement by adjusting our baseline treatment in controlled ways and, all the while, evaluating the efficacy of each adjustment. These kinds of experiments are conducted by scientists, engineers, and statisticians in the lab and field with tools such as Excel, SpotFire, proprietary simulation software, and various programming languages. However, the full power of data is just beginning to be leveraged industry-wide for two reasons:
- Powerful datasets take time to accumulate.
- The emergence of open-source tools (like Python and R) have leveled the playing field for advanced analytics among startups and supermajors alike.