You are Home   »   News   »   View Article

Agile Data Decisions - new company by ex CGG Data Management Staff

Tuesday, March 21, 2017

A new data analytics company specialising in unstructured data has been set up by three former senior managers of CGG Data Management.

A new data analytics company, focusing on understanding unstructured data (such as documents), has been set up by three former senior employees of CGG Data Management.

The company aims to develop and commercialise technology initially developed by CGG Data Management, which CGG is no longer interested in developing. The new start-up has inherited of the full ownership of the technology and will continue to support CGG using and commercialising it.

Jacques Micaelli, IT leader and CEO of Agile Data Decisions, was formerly VP Technology IT with CGG Data Management Services. Amit Juneja, co-founder of Agile Data Decisions, was formerly the 'Premier Data Scientist' with CGG Data Management, and has 16 years of experience in machine learning. Henri Blondelle, the third co-founder and VP sales and marketing, was formerly VP global business development with CGG Data Management, and is a trained geologist.

The company has hired three full time software developers working in New Delhi, and is also supported by some independent advisors, including Philip Neri, formerly VP global marketing with Ikon Science.

The company is registered in the US because both Mr Micaelli and Mr Juneja are based there, but it will probably be later registered in Europe as well, Mr Blondelle says.

The company is seeking $500k to $1m investment before the end of 2016 to enable it to industrialise the prototype, and then seeking between two and four reference clients before the second quarter of 2017, Mr Blondelle says.

This could lead to a second round of funding in 2018. The technology can work on data other than oil and gas data, but 'we are focusing on oil and gas for the time being,' he says.

Unstructured data

The mission of the company is to find ways to bring analytics to the 80 per cent of subsurface data which is unstructured.

As an example of how it might work, consider that a venture capitalist might receive hundreds of different written proposals, and want an automated way to classify them by a criteria which makes sense to the venture capitalist - such as which market the proposed company is in, and what level of funding it is looking for.

To illustrate the value, consider that legacy car companies have enormous amounts of data, mainly stored in (unstructured) written reports. They have modern competitors like Tesla, who (probably) keep all of the data in structured format, so it is much easier to analyse.

In order to compete, the legacy car companies need to be as good at analysing their unstructured data as the modern companies are at analysing the new structured information, Mr Blondelle says.

Workflow

The workflow starts with extracting text from the documents (using optical character recognition).. The next stage, where Agile Data Decisions adds value, is running algorithms and machine learning to try to get some understanding of what the text is about and index it to a category.

This is then used to give the document automatic metadata. Finally, the resulting extracted information will integrate with analytic tools, such as Spotfire and Tableau, which will analyse the data, Mr Blondelle says.

Agile Data Decisions aims to do the same task on unstructured data as a human data manager would. Data managers develop an understanding of the various data objects common in oil and gas exploration and roduction, such as 'composite logs', well reports, VSP, core reports, and how to recognise where the value in them is, Mr Blondelle says.

Data managers also learn how to index documents, for example by well and by category. Data managers might also learn to spot the difference between good and bad data, and assign a confidence value.

Similarly, Agile Data Decisions can associate a confidence value to the data it extracts. The software can also create automatic links between data and documents.

For example, if the analysis shows that a certain depth parameter has three different values and they should all be the same, it can automatically link to the original documents, so someone can make an assessment of which value is most credible.

Competitions

Many oil and gas companies and government departments have recognised that there could be enormous value in their unstructured data, and have organised competitions to select a company which can help them realise this value.

In May 2016, the organisation Common Data Access (part of the oil and gas UK national association Oil and Gas UK) announced an 'unstructured data challenge', where it would work with a number of data and analytics tool vendors, to see how good their technology is in making sense of the archive of documentation describing over 11,000 wells and 2,000 seismic surveys, and extract value from them.

From the project, Common Data Access hopes to be able to demonstrate the value that modern data and information analytics tools can add to a large corporate collection of unstructured subsurface data. Agile Data Decisions is submitting its technology to this challenge, Mr Blondelle says.



Associated Companies
» Agile Data Decisions
comments powered by Disqus

CREATE A FREE MEMBERSHIP

To attend our free events, receive our newsletter, and receive the free colour Digital Energy Journal.

DIGITAL ENERGY JOURNAL

Latest Edition Aug-Sept 2020
Aug 2020

Download latest and back issues

COMPANIES SUPPORTING ONE OR MORE DIGITAL ENERGY JOURNAL EVENTS INCLUDE

Learn more about supporting Digital Energy Journal