By Sarah Byrne
COMPANIES are turning to data scientists for data analysis, in a bid to optimise operations and reduce risk at a time of lower gas prices, Kaggle chief executive and founder Anthony Goldbloom says.
Speaking with Oil and Gas Australia, Mr Goldbloom said Kaggle works with four out of the 20 biggest producers in North America and Canada, two of which are Shell and Apache.
Mr Goldbloom said data driven models were the future of the industry.
“There has been a growing recognition of how analysing data can contribute to well optimisation and reduce risk during land acquisition,” he said.
Data scientists come together through Kaggle to solve problems for companies through a competition forum on the website.
The company keeps an eye out for top performers in these competitions and often offers them an in-house role as a data scientist, analysing mass amounts of data for oil and gas companies.
A problem faced by oil and gas companies is finding a way to interpret the vast amount of data during land acquisition and from well logs, Mr Goldbloom said.
Mr Goldbloom said one of the oil and gas companies he has worked with said “The data is screaming at us but we don’t know how to listen to it”.
With the price of oil continuing to fall, companies were looking for cost effective ways to optimise their existing operations.
Data scientists at Kaggle analyse public and private data supplied by oil and gas companies to make recommendations on how to improve efficiency on sites.
Mr Goldbloom said in one case the company assisted an oil and gas producer by analysing data from well logs and providing recommendations of how much fluid should be pumped into wells for hydraulic fracturing to optimise its operations.
Land acquisition was one example provided by Kaggle as a way analysing data can reduce risk and provide explorers with confidence in their acreage, something sought after in the current economic climate.
“One customer thought a particular core measurement showed whether the land was going to be good for production or not. By analysing the data we found there was very little link between the core measurement and good production,” Mr Goldbloom said.
“Geology would tell you the core measurement is important, but our data scientists found the measurement wasn’t consistent and gave varied results,” he said.
Data scientists at Kaggle analyse data sets to calculate based on the data how much commodity the company may be able to recover from a particular area of land.
Mr Goldbloom said working in this way reduces risk and provides oil and gas executives with confidence when deciding where to drill or making operational changes on site.
Kaggle was founded by Anthony Goldbloom in 2010, it is an online community of data scientists, including a team of in-house data scientists and is based in San Francisco.