This talk was presented at Saga Wisdom SÖKKVABEKKR in April 2022. It's not a technical talk, but a mix of memoir and manifesto which tries to piece together some lessons from history to make sense of important questions like “what happened to the ‘digital engineer’?” and “why is it so rare for data science projects to succeed in the modern oil and gas industry?”.
This talk was presented to the Houston chapter of the Society of Petroleum Evaluation Engineers (SPEE). It's about the three big ways in which stochastic simulation (Monte Carlo methods, for short) can fail in practice in common reservoir engineering and petroleum economics applications. It also resulted in the creation of some great software, which we're hoping to offer commercially in the near future. The in-person talk largely consistent of an interactive demo, but we've tried to capture some narration and screenshots in this “Director's Cut” slide deck. (Also, this talk has a cute baby.)
This talk was presented for the 2021 SPE Geothermal Datathon. It was pitched to us as a 40-minute talk about storytelling with data; we later found out it was pitched to the participants as a “bootcamp”. Sorry, kids! In hindsight, I think something more HBR-approved (or perhaps the sort of thing which appears on cutting-edge Medium blogs) might have gone over better, which is too bad, because we were trying to convey a message we think is important.
In this almost entirely non-technical talk, we receive no commission for promoting Edward Tufte's books, quote Slavoj Žižek and Richard Feynman on ideology and science, and generally try to argue for data scientists to focus on educating rather than persuading in their technical communication.
We also use some photographs from a trip to Las Vegas to point out common fallacies—or underhanded tricks—in data storytelling.
This talk, originally presented as a webinar hosted by Saga Wisdom and then greatly expanded for a webinar hosted by the SPE Permian Basin Data Analytics Study Group, slowly and gently weaves its way to the thesis that programming is the keystone skill for a petro-technical professional interested in data science or machine learning. Many of the thoughts and insights that eventually crystallized into our Practical Programming for Engineers course can be seen in a more raw form here.
We review the history of “gatekeeping” in computer science education, deconstruct some jargon, and build some bridges from the skills engineers already have to the world of computer programming.
Case studies are presented by both Terminus founders from the oil and gas and chemical industries.
We conclude, lamely, by recycling slides from the “Found in Translation” talk.
This technical talk, presented for several audiences in 2017, introduces the concept of "domain-specific languages" to a non-specialist audience. We discuss the general idea of "notation as a tool for thought", and introduce an actual commercial project as case study, in which a custom domain-specific language adds critically-required flexibility to an economic evaluation system, at a low cost in complexity.
Unfortunately, no recording is available, but slides and example program code are freely available on GitHub.
This short talk, presented here as an interactive slide deck with audio narration, attempts to deconstruct the state of data science in the oil and gas industry circa 2015.
We also present a few bits of advice for data science practitioners working in the industry.
Mr. Turk is the primary author of SPE 163713-MS (“Analytics Beyond R2: Year One”) and has presented at multiple industry events including the 2013 SPE Digital Energy Conference. His focus is on finding new and innovative ways to get insights from data science into the hands of the technical professionals making business decisions, including the deployment of creative and interactive visualization tools.