Lars Pittman

Oil is the New Data

On the 25th March the Oil and Gas authority (OGA) in the UK released over 130 terabytes of data from the UK Continental Shelf (UKCS). The data, covering more than 12,500 wells, 3,000 pipelines and 5,000 seismic surveys, is now publicly available from the Oil & Gas National Data Repository (NDR).

Momentum for use of data science and artificial intelligence in oil and gas is building, and with releases such as the Equinor Volve data set in 2018, the industry is adding fuel to the fire with a hope to further accelerate the development of insight and technology. The data set is released on the OGA’s already established open data centre which also includes production data as well as several interactive maps and dashboards.

A brief exploration of the data

Accessing the NDR data is straightforward by registering as a user and logging in. The data is organised through interactive maps and tables. The maps provide a neat way of viewing infrastructure, seismic coverage and well-locations. With the data you can of course plot any visualisation yourself. Here we show an example of well locations on the UKCS plotted in a Jupyter Python notebook using GeoPandas and Matplotlib.

Location of wells on the UKCS split into permanently abandoned (OGA well status code AB3) wells and active wells (including temporarily abandoned wells, OGA well status codes: AB1 and AB2)

For those of you that are interested in the code, you can see the full notebook here.

The well data includes columns such as ‘spud date’ (when the drilling of the well started), ‘well type’ and ‘status’. Since 1964 a total of 12,157 well-bores have been drilled on the UKCS, only 5,173 of these are permanently abandoned (ie. permanently made safe and inaccessible). The remaining, almost 7,000, are still active or only temporarily abandoned.

Histogram of age of wells on the UKCS; permanently abandoned (AB3) wells in orange, and wells still active or temporarily abandoned in blue

The peak activity on the UK continental shelf was in the mid to end 1990s when around 700 well-bores were drilled annually. Since then, the activity has dropped to just under 200 spudded wells in 2018.

At the end of 2017 the reported 2P reserves on the UKCS was 5.4 billion barrels of oil equivalent (boe), equivalent of 20 years production. The OGA is hoping that the release of this data can help halt the steep decline in well activity and drive investment to unlock what it claims to be an estimated further 20 million boe from the shelf.

Conclusion

The possibilities for use of this data set is wide ranging and will certainly be of great value to academia and industry alike. Collaboration and ability to share date freely will encourage publication. At Taank Technologies we are always glad to see more data being released publicly as this can only benefit data science and the industries it supports.

by Lars Pittman & Cammil Taank

ctaank

How to make the best use of Data Science

Data Science is a rapidly evolving discipline, and seemingly a must have component of any sizeable business. It’s clear and widely agreed that more data, and better understanding of that data leads to better business decisions, but most organisations do not have the expertise to understand how use data science effectively.

The aim of a successful data science function is to maximise business value as efficiently as possible. Algorithms are important, but they are only one aspect of this problem.

Alignment with business goals, efficiency of teams and projects, ability to act upon results of new algorithms are all key to maximising the value derived from data science work.

The key then to maximising value is firstly to understand the problem that needs to be solved. This problem is rarely a purely mathematical problem, but rather a mathematical problem combined with the business changes that can utilise the mathematical solution effectively.

We urge you to think about this carefully before embarking on data science work – it is the single biggest source of wasted resources.