Data Without AI is as Useful as Sunglasses on a Dog
Data Science & Analytics

Data Without AI is as Useful as Sunglasses on a Dog

Rosemary Jackson  •  

Should dogs wear sunglasses: No. Maybe older dogs with sensitive eyes or cataracts, but it’s generally agreed that except for a fabulous fashion statement, sunglasses don’t do anything for dogs.

Are AI data driven answers the best: Yes. Targeted data loaded and applied in AI (artificial intelligence) driven algorithms provide tradeoff solutions to specific questions. BI (business intelligent) data analytics can build some lovely graphs, but the application to solutions is not always relevant or clear. AI and data together “drive ever more accurate, complex, and sophisticated predictions.” Without AI, your data isn’t doing enough for you.

There’s only one good thing about data. When it works for you. If your data is sitting in some repository, if your data is not in AI, if your data is in an excel spreadsheet waiting for you to integrate it into some data analytics business intelligence platform, then your data is as useful as sunglasses on a dog. It may look interesting, but it’s not doing much.

Your data needs to help you answer the ‘what if’ questions. Engineers and scientists ask tradeoff questions,” Dr. Nitin Chaudhary explains, “but they wind up spending more time in data integration and data warehouse. How much time does a reservoir engineer spend on what he cares about most? We invert that and take care of all the lower layers. We give your team a chance to focus on reservoir management, for geos the well placement and the geo modeling.

“In answering those tradeoff questions, you need to be intimately tied to the data you’re loading in and how you’re conditioning it to answer that question. We’ve created the best ‘what if’ experience you could ever imagine having.

At we have an automated data pipeline. For our customers, we provide daily updates for their own wells. This includes updating a well's meta information, daily production. For the non-company wells (public data), we do a weekly auto update connecting to a vendor data-source. All of this happens with no intervention from either side. This kind of auto-update feeds the productivity models inside and helps the operator keep tabs on the productivity of their internal wells as well as what the competitors are doing and improving.”

Charles Connell, VP of Product, continues, “Updating depends on the data type. Something like microseismic you might only input once a year. As Nitin indicated, we’ll bring in production data which could be daily production or monthly production so those might update on a daily or monthly cadence. Other data types like the completions data we would bring in that pump data every time you complete a well. If you have one crew running that might be monthly. For larger clients, who have several completion crews running, that could be a weekly data load.

“ has done this enough now that we know what data is needed. If a client doesn’t have any data, we can still get started with publicly available data in some of the models we’ve already built. We don’t need their data to get started, but we are a data driven analysis so the more data we can get the more granular we can make the analysis. If a client doesn’t have microseismic data that’s fine, we can still do the analysis. If they do have that data, then we can tune the models and adjust it and bring that in and account for some specific things they’ve seen.”

Usually, we manage loading a client’s data into our standard file formats that we process and integrate into a common data structure. Certainly, the system is set up so our client companies can use the import wizard and upload the data. What we’re finding is that more of our clients prefer just to get the answers and the reports. They send us the data; we upload it and generate the post job report. Or we set up a data base connection where we bring in their monthly production. Every month we have a script that’s running, and we bring that data in. It can be on the end user, but usually it will be something we set up to automate or we do ourselves.

Different basins, different geologies, different size pads create different tradeoff questions which might require us to make something new that becomes part of a client’s exclusive platform. If the customer company has more completions data, more microseismic, more logs, complements what we can already bring in, in terms of our models, with their data.” understands that a different or new question that’s part of a company’s understanding of the subsurface necessitates a new set of tradeoff answers. That will mean bringing in additional data or that we have a different computation to create those answers. New questions, additional algorithms, specific data, ingests and automates so that you’re not wasting your time putting sunglasses on a dog. Your data is producing real results and real options that supply the information you really need now.

114 Main Street Ste. 400
Houston, TX 77002
4-5609 Avenue du Parc
Montreal QC, H2V 4S8