In his role as Verdazo Analytics’ Senior Technical Machine Learning Advisor, Tyler made impactful contributions to the Calgary section of the SPE, its members, and our industry.

We invite you to join us in celebrating Tyler’s accomplishments at the upcoming Data Analytics Breakfast event at the Calgary Petroleum Club on January 10th from 7:30 to 9:30 a.m.

Here is just a small sampling of Tyler’s notable contributions:
Machine Learning: Practical Use in Upstream Oil & Gas
Machine Learning: Is it really a Black Box?
Machine Learning: Finding the signal or fitting the noise?

Since launching the division in January 2018, Verdazo Analytics has delivered consulting services to clients and producers, and in partnership with other industry-leading consulting firms. The company has helped the Oil & Gas industry understand and realize the practical benefits offered by Machine Learning in the areas of production performance prediction and completion optimization.

The expanded team joins Tyler Schlosser, the company’s Senior Technical Advisor for Machine Learning and Reservoir Engineering. Together, they’ll work on developing Verdazo Analytics’ forthcoming Machine Learning software and helping clients understand and adopt machine learning technologies and methodologies.

“A larger, diverse team allows us to respond to the growing interest in leveraging complex datasets to produce valuable insights that contribute to the bottom line” says Verdazo Analytics President Bertrand Groulx. “The success we have enjoyed in our most recent consulting projects has allowed us to craft our technology and processes to extract insights from a broader collection of data.”

The new staff bring extensive experience and education across geophysics, geomechanics, reservoir engineering, completions design, drilling, infill programs, enhanced recovery, economics, math, physics and software engineering. Many of the new team members also have recent experience using machine learning techniques to tackle some of the most complex problems at major producers.

“Our team has domain expertise, knowledge of Oil & Gas data sources, and deep machine learning expertise,” says Groulx. “This combined skillset is compelling on its own, but when you empower the team with the VERDAZO visual analytics software platform for data integration and visualization – that’s when things really get interesting.”

Verdazo Analytics Machine Learning Division

To learn more about Verdazo’s Machine Learning capabilities, contact Dean Anderson.

Groulx will be joined by speakers Bob Lamond, VP Exploration, Rife Resources, Ltd, and Darcy Fairbrother, Strategic Sales Project Manager for NCS Multistage, Inc. The group will walk through corporate case studies and there will be time for audience Q&A.

The panel is part of the latest session put on by the Canadian Society for Unconventional Resources’ (CSUR) focusing on Data and Data Analytics for the Upstream Oil & Gas Sector.

Register for the event and learn more here.

The theme of this year’s summit is disrupting the concept of closed networks. Attendees will learn about the value, support and flexibility in technology interoperability driving operational and financial performance.

Groulx is joined on the panel by Elena Dumitrascu (TerraHub), Téo Adams (ATB Financial), Bryan Pederson (Amalto Technologies) and Daniel King (EnergyNow).

Register for the event or learn more here.

ATCE 2018 will feature 350+ technical presentations, special sessions, and networking events for over 8,000 Society of Petroleum Engineer members and other professionals looking for the latest news on current and future technologies to help find and produce hydrocarbons faster, more efficiently, and safely. The conference takes place in Dallas, Texas from September 24-26, 2018.

Learn more about ATCE 2018 and register here.

Consider starting with the data you have, take it for a test drive. See how the business wants to use it and evolve your data quality and architectural initiatives incrementally. You’ll realize value on the way and better focus your efforts.

Why do they fail?

1. Lack of focus: Hyped up terms like “big data”, “data lakes” and “cloud” distract us from the pragmatic task of delivering information ­­­— getting reliable, current information into the hands of business users in a form that allows them to inform decisions. Don’t over complicate the task at hand.
2. Slow time to value: While these initiatives are inspired by sound motives, they fail to get buy-in or build momentum because they aren’t delivering near term value. Successful projects are built on a series of quick wins and can make course corrections along the way to ensure success.

Context for success

I like the “cart before the horse” analogy because I think of the cart as the data, the horse as the visual analytics tool(s) and the driver as the business user who is using the data to make value-based decisions. Like the horse-drawn cart has a destination, so should the data-driven decisions — fulfilling corporate objectives.

The tools, to some degree, don’t matter as long as they provide reliable information delivery in a form that supports decision making processes. And yes, there needs to be a process. In my blog post Innovation or Dishwashing Robots I discussed how companies often try to apply new technology to an old way of doing things. Time to value should be the primary driver, and adapting processes to leverage the strengths of tools yields faster, better results than trying to adapt tools to fit legacy processes. The goal should be to use technology to deliver consistent, reliable information to inform decisions that are aligned with corporate objectives (outlined in the 2017 SPE talk Is your organization competing on analytics? A roadmap to analytics success).

Tips to be successful

• Make Time-to-Value your primary focus: Build momentum with small successes. Get something useful into the hands of business users sooner…don’t over-architect the solution because no one can anticipate all of the needs until they walk down the path.
• Select visual analytics tools with these qualities:
  1. Data Governance –there should be administrative capabilities to ensure that data is delivered with the same business logic for all users. Consistent, reliable and repeatable should be the information delivery objectives.
  2. Adaptability & Agility – your data needs are going to evolve over time, as might your data architecture. Having a tool that can be configured quickly to evolve with your needs is important. This should include the ability to change data mappings as your data sources evolve (e.g., from a direct database connection to a data mart) without compromising end-user analyses. The consumption of reliable data, wherever it resides, should be seamless to the end user.
  3. Usability – you want maximum adoption. Aim for optimal usability but don’t get caught in the trap of designing something unique for every user…try to standardize on some best practices and processes in your organization and don’t forget to train your people.
  4. Self-Service – you want people to have some latitude to build their own analyses with governed data in a sanctioned environment…consider it to be “governed freedom”.
  5. Domain-Specific – domain-specific tools often have industry expertise and best practices baked into them, and specialized processing capabilities that are difficult to accommodate by pre-conditioning the data. For example, time-normalizing data could take many forms (e.g., first gas production, first oil production, peak production date, spud date, etc.). Trying to accommodate these in a data warehouse with pre-conditioned data would be extremely challenging, time consuming and expensive. Domain-specific tools often deliver a much faster time to value and can be more cost effective versus building and maintaining a build-it-yourself solution. For a more in-depth perspective, read Should you build or buy software for your company?
• Define and Evolve Processes: Explore the opportunities for ideal business processes using your evolving technology, identify best practices and try to standardize the approaches that are most effective. I have a client that regularly showcases successful innovations in analyses and processes in an attempt to fuel a culture of excellence. Let the rock stars share their techniques and innovations with others so everyone can improve. You also need managers who are equipped to lead and establish the standards and processes that are aligned with corporate objectives. It’s their job to establish expectations and communicate them clearly.
• Training: Don’t forget to build training into your budget. It takes people, process and technology working in concert to optimize efficiencies. Training, like technology, should be perceived as an information and insight investment. People are driving technology to make your company more successful…empower them with adequate and regular training. Training that integrates with your company’s processes is even more impactful.
• Business & IT Collaboration: IT is there to support the business and should be grease in the gears, not sand. The best way to build a successful project is to have business and IT collaborate with a clear understanding of mutual goals centered on creating value in a timely fashion.
• Executive Sponsorship: You need the support of the people who sign the cheques. They’re more likely to buy into an approach that has time-to-value at the forefront and is conveyed as an investment in insight, efficiencies and optimization. Quick wins that are clearly defined will cement executive support and make everyone look good. It allows you to wade into the waters of pragmatic spending and justify the cost of each step rather than strive for a massive budget with lofty goals that use hype-inspired terminology.

Avoid distractions

I use the term “magpie syndrome” to characterize someone who is always chasing the shiny object at the expense of getting a solid information delivery foundation in place. There are no shortage of visually impressive charts, dashboards and AI/machine learning technologies available to you…but beware of diverting your focus away from the core principles of effective information delivery ­­­— consistent and reliable data that supports value-based decisions that are aligned with corporate goals. Trying to cram more information into a dashboard or fussing endlessly about colours and layout are things that will take your eye off the prize. While these can be important considerations for clear interpretation and insight, they fundamentally rely on clear goals and the right data. Experts in the field say that several individual charts usually convey insight more effectively than a complex dashboard. Don’t make things more complex than they need to be. Ask yourself repeatedly “is what I am doing helping me make a better decision?”.

In summary, don’t start with the data, start with deliverability. Make time to value your focus and get proper executive support. Be agile, be nimble and be prepared to evolve. Processes should be well-defined and supportive of corporate objectives. Failure to put some structure around data governance and business processes will lead to chaos…empower your team with “governed freedom”. Get started – your first success should only take a few weeks. If it takes longer, the success of your project may be at risk.

There’s an enormous amount of discussion about the possibilities of Machine Learning but far less practical information about the impacts it can have on Oil & Gas today. In this presentation, we articulate some of most valuable prizes Machine Learning offers for upstream Oil & Gas, including feature importance and sensitivity analysis, and the power of predictive Machine Learning models. The presentation outlines two recent case studies – one focused on reservoir properties and the other focused on drilling and completion optimization – that bring the material to life. This presentation was originally delivered May 31, 2018 as part of the SPE Oil & Gas Breakfast Series in Calgary.

What is a “Black Box”?

Machine learning models are sometimes characterized as being Black Boxes due to their powerful ability to model complex relationships between inputs and outputs without being accompanied by a tidy, intuitive description of how exactly they do this. A “Black Box” is “a device, system or object which can be viewed in terms of inputs and outputs without any knowledge of its internal workings” (Source: Wikipedia).

Black Boxes (and Machine Learning models) exist everywhere

We tend to label things as “Black Boxes” when we don’t trust them more than when we don’t understand them. Machine Learning models aren’t unique in having an element of “mystery” in how they work – there are all sorts of things we trust all around us for which we don’t fully understand the inner workings. GPS, search engines, car engines, step counters, even the curve fitting algorithms in Excel are examples where we trust what’s happening inside because we’re able to see and, with experience, have confidence in the results they produce.

Machine Learning itself is everywhere, and is already “trusted’ widely by nearly everyone, whether we realize it or not. Google Maps, airline autopilots, spam filters, optical character recognition (OCR), shopping recommendations, depositing a cheque at an ATM, fraud protection, voice-to-text on our mobile phones, and even some best-in-class medical diagnostic techniques are all things that rely on Machine Learning to be as effective as they are.

(Supervised) Machine Learning is just another fitting algorithm

Most of the time, when people talk about Machine Learning in oil and gas, they are referring to Supervised Machine Learning. This is where an algorithm learns by example after observing many cases of “given data X1, X2, X3 and X4, the outcome was Y”. With enough examples, an algorithm can learn how to predict the outcome from this type of data, assuming a relationship exists.

This is basically the same process as fitting a linear model or calculating a linear correlation coefficient in Excel or other tools, but with the ability to handle more complex relationships. Linear regression is included as part of the “Machine Learning” algorithm toolkit by all prominent open source tools. Sometimes, that’s all we need. Other times, it’s useful to have more powerful tools that can better handle many inputs and the complex, nonlinear relationships they may have with both the outcome we’re trying to predict and with each other.

Affords us unforeseen insights

The most obvious way Machine Learning differentiates itself from a Black Box is its powerful ability to describe feature importance. Feature importance is a measure of how useful a particular input (or “feature”) is in predicting the outcome. The power of Machine Learning can make it difficult to grasp all that’s going on under the hood. This same power gives us a very high level of confidence that if there is a meaningful relationship to be found, Machine Learning will find it. However, this power should be used carefully as Machine Learning can produce results that could be misleading (see our blog on Finding the Signal or Fitting the Noise?).

Conversely, if there is no predictive relationship found between a particular feature and the outcome, we can be confident that none exists. Gaining an understanding of what matters and what doesn’t in making predictions, and to what extent, is very valuable in better understanding the problem.

Test hypotheses quickly

The potent ability of Machine Learning to discover and communicate relationships in the data make it an excellent tool for hypothesis testing. Without Machine Learning the task of investigating a hunch like “I think horizontal well inclination and frac intensity together can tell me a lot about how much gas a well will produce in the first year” can be time consuming. It can even yield a false conclusion if the relationship is complex or other influencing factors aren’t properly accounted for. Sometimes relationships are nonlinear, threshold-driven, or exist among interactions between different types of data. Machine Learning excels at uncovering complexity. An underappreciated benefit of Machine Learning is that it allows us to focus further attention on only the most promising leads and avoid going down dead-end rabbit holes.

Visualization brings transparency to Machine Learning

We use interactive visualizations in VERDAZO throughout our Machine Learning process to add transparency, identify data opportunities and steer the process toward better results. Visualizations that we typically use include:

• Statistical and map views of the data to help select sample data
• Feature importance (tornado plots)
• Comparing Model results in cross-plots & cumulative probability distributions
• Fitness assessment and error characterization using maps, distributions and probit plots, slicing and dicing by categories, vintages and quartiles
• Sensitivity profiles (applying variations to one value while holding all others constant)
• Parallel coordinates distributions for validation

In the end, Machine Learning isn’t the Black Box that many perceive it to be, especially when it involves a rigorous process and is supported by a robust set of visualizations. Our Machine Learning services are purpose-built to help our clients get the greatest clarity, understanding and insight from their data – and build trust in the reliability of the results.

Groulx will share the podium with co-presenter Jim Gouveia of Rose & Associates. Their course will provide a solid background for professionals in the application of Statistics, Uncertainty and Aggregation Principles to their evaluations. At the end of this course, participants should have an intermediate understanding of estimating under uncertainty, basic statistics & aggregation principles, advanced analytic methods and selected best practices in production type well curve development.

Groulx and Gouveia’s course takes place on Saturday, June 2 and runs the full day from 8 a.m. to 5 p.m.

Learn more about the conference agenda and register here.