Introducing Novi Data Network: Fueling Machine Learning Forecasting with Proprietary Data

Novi Data Network (“NDN”): Leveraging Proprietary Oil and Gas Data for AI Driven Well Forecasting Models.

At Novi Labs, we believe high quality well data processed by industry tuned algorithms is the key to the most accurate data driven models.. That’s why we are excited to announce Novi Data Network (“NDN”), enabling Novi customers to pool and utilize proprietary production and completions data for training machine learning models.

Leveraging Novi Data Network (“NDN”) data is effortless –  it is fully integrated into Novi’s end-to-end oil & gas analytics software workflows. By utilizing Novi’s industry-specific patented machine learning platform with NDN’s unparalleled dataset quality, customers can confidently develop forecasting models that enable better business decisions.

With over 20,000 wells from more than 50 operators, NDN sets itself apart by offering its members exclusive access to proprietary data that is not typically disclosed to regulatory agencies, including non-allocated daily production, completion stage and cluster counts, target formations, and more.

NDN has a key difference with a standard data consortium: NDN data is not exposed or downloadable by NDN members. The data is used only for training models, which allows the models learn from a pool of information without actually sharing the data. In order to access NDN, you must be a Novi customer and an opted in NDN member, and must be shared on a regular basis to Novi so that NDN is kept constantly updated.

Why is Novi Data Network needed?

Even the best public data has its limitations! Here at Novi, we put tremendous effort into cleaning and upgrading data from public sources, like the Texas Railroad Commission or FracFocus. But even with rigorous data processing and preparation, public data still has challenges, which vary depending on the jurisdiction. In Texas, for instance, production is reported at the lease level (not well level), so providers like Novi allocate the lease level production down to the well level. In Alberta, the liquids content of gas wells is not usually reported. In Argentina, bottom-hole locations aren’t public (good luck with spacing calculations!).

What does this all mean? If you are an operator, you likely have a quality dataset with all the information you would need for your own wells, but learning from offset designs can be tricky where the production data has noise from allocation, unknown downtime, and uncertain completions. But those offsets probably provide the most valuable information to you precisely because their designs are different!

So what can you do? Of course, licensing public data from a company like Novi is a great step, but joining Novi Data Network gives you access to much more proprietary data.

How does the Novi Data Network add value?

1- NDN for Development Planning

For development planning, NDN proves invaluable, especially in states like Texas where public completions data is typically limited to proppant and fluid volumes. NDN grants access to stage and cluster designs from dozens of operators, allowing for the identification of opportunities to improve production, decrease costs, or achieve both. For spacing and stacking optimization, accurate production data is essential, particularly for large multi-well leases with intricate parent-child relationships, where allocation proves most challenging.

2- NDN for Acquisitions & Divestitures

Business development teams can benefit significantly from the NDN when venturing beyond their core area. NDN provides a proprietary dataset for A&D valuations, enabling the generation of highly accurate forecasts and the discovery of optimization opportunities, even in unfamiliar basins. With NDN, business development teams can maintain their competitive edge and make well-informed decisions in new areas.

3- NDN for Planning and Budgeting

Novi Data Network (NDN) plays a crucial role in budget and planning processes by providing accurate well-level data for model training purposes, which is essential for forecasting development designs and upcoming wells. Traditional issues, such as lease-to-well allocation, can hinder forecast accuracy, but NDN mitigates this concern. Additionally, NDN is built on daily data that allows for precise downtime removal in training data, addressing unknown downtime’s impact on forecast accuracy. In jurisdictions like Alberta, where liquids content of production is often undisclosed, NDN enables the training of models with proprietary liquids production data, further enhancing accuracy.

How does it work?

NDN includes the following attributes:

• Daily, well-level production for oil/gas/condensate/water
• Well surface hole and bottom hole locations
• Target formation
• Proppant, fluid volumes
• Total clusters
• Total stages
• Lateral Length

Beyond that, all NDN data have been cleaned, processed, and reviewed by Novi reservoir engineers. These steps are above and beyond what our operator customers have already completed; our goal is simply the highest quality data in the industry.

Within Novi’s analytics platform, using NDN data is easy. You can add in NDN data just like you would an Insight Engine data source, then configure it in collections, data sets, and models. In order to preserve the privacy of the NDN data, we have restrictions on outputs, but we still provide summarized statistics and anonymized results (such as SHAP values) to help you interrogate model performance.

How is NDN different from Novi’s Proprietary Production Data?

You might be already familiar with the proprietary production data that Novi integrates into its Insight Engine offering. There are a couple keep differences that I want to highlight here:

• Proprietary production data IS shared with all our customers; NDN is only for members, and it is not shared but rather pooled for model training.
• Proprietary production data is monthly, well-level cumulative production volumes, whereas NDN includes daily, well-level production plus all the other data mentioned above like clusters and stages.
• NDN can only be leveraged with Novi’s analytics platform (Data Engine, Model Engine, and Forecast Engine), whereas we provide the proprietary production data to all our customers.

Beyond simply offering accurate forecasts, we have built transparency into our models, developed a full self-service platform, and integrated public datasets into our applications. Proprietary datasets like our production data offering and the Novi Data Network are just the latest step in nearly a decade of pushing the envelope in advanced analytics for oil and gas.

Have any questions regarding Novi Data Network (“NDN”)?

You can always request a free demo to see what’s under the hood and how Novi Labs can help you to unlock hidden investment opportunities.

Ted Cross

Ted Cross is VP of Product Management at Novi, where he builds powerful machine learning data and analytics products to solve the toughest problems in unconventional oil and gas. Prior to Novi, Ted was a Senior Geologist with ConocoPhillips.