Energising data science
There are few industries that have remained completely unaffected by data science; its applications are simply too widespread. Industries that previously did not have a data driven approach now have the opportunity to make up for lost time. Although Woodhurst’s primary focus is Financial Services, the energy sector provides some great examples of how data science and machine learning are helping to change the landscape of this industry.
Finding oil is an inefficient process – it takes a lot of trial and error to find a well with a sufficient amount of oil to drill. Now, big tech companies such as Google, Amazon and Microsoft, are teaming up with oil giants to offer their expertise in an area of machine learning called ‘Deep Learning’. Deep Learning is facilitated by Artificial Neural Networks (ANNs), which is a connected system loosely inspired by how the brain works. This is a quickly evolving area with many applications.
Tech companies are creating Deep Learning ANNs to more easily detect key geological features that indicate the presence of oil. This more efficient and accurate process will lead to billions of pounds of savings.
One could argue whether the use of technology to support the depletion of the planet’s natural resources is ethically or morally correct, but the technology is also being employed to ensure individuals and business reduce and improve their energy usage.
Reducing energy consumption
Much of the machine learning focus in the energy industry is on how it can be implemented to identify opportunities to reduce consumption and improve efficiencies in production. According to The Association for Decentralised Energy, a quite incredible 54% of electrical energy created in the UK is wasted, which means that data science can have a huge impact on energy savings. We can use a process called demand management to understand where waste is occurring, identify how we can reduce it, and better serve a lower, more efficient energy demand. According to the National Institute for Science and Technology for the US, by 2030 data analysis and science of energy production could generate savings of up to $2 trillion.
But how can data science actually help reduce energy waste? In the home, technologies like smart thermometers can feedback data into an analytics engine which is then used to reduce unnecessary consumption and lower energy bills. This data, aggregated and combined with other sources, can also be very important in creating accurate machine learning models to predict the energy demand for areas of the country at any given time. This means that energy producers can produce to meet these demands, limiting any waste.
On an enterprise scale, Google have used AI to reduce its Data Centre Cooling Bill by 40% and produced 15% less energy in 2018 than 2017. Microsoft have launched the start-up focused ‘AI Factory’. Its aim is to solve give major economic and societal issues in Europe, one of which is a climate change initiative that incorporates energy usage.
Considering the cost reductions that can be achieve, and the increase in availability of AI-enabled demand management, it is not unreasonable to think that soon all businesses and homes will adopt it.
Conceptually, you can begin to consider how Financial Services might similarly use AI to put its scarce resource – capital – to more efficient and effective use.
The Great Crew Change
The oil industry is facing a problem it calls ‘The Great Crew Change’. Many of the baby boomer generation are retiring – in 2015, 50% of the work force were due to retire in the next 7 years. Unsurprisingly, this has led to the pervading belief that a growing skills shortage is by far the biggest challenge the industry faces.
In situations like this, data science and machine learning can be used to pick up the slack – particularly in operationally heavy workloads.
It has also been noted that using a technology orientated process can be very effective at attracting young graduates straight out university, which the oil industry often struggles to do. Between 2013 and 2018, the number of graduates entering the industry halved.
Although the skills challenge in FS is less stark, there is the shared need to attract and retain technologically-savvy people that can embrace these new technologies and allow them to thrive.
AI can have a major transformational impact in improving business practices, but it is important not to forget about cases where AI can make small cumulative differences.
In the energy sector there are many areas where data science can improve efficiencies by a couple of percentage points in lots of different areas. For example, the concept of survival analysis can be implemented. In medicine, this is a statistical method that estimates survival rates for patients based on their condition, treatments and other relevant factors. In the oil and gas sector, this concept has been applied to field equipment to predict when it will need replacing.
Data science has also been used to predict and detect outages. Previously, engineers used static algorithms and models, but now they have been able to upgrade to data driven real-time solutions.
There are dozens of examples where data driven models are making a small difference, but cumulatively they really add up. Once the enterprise is set up to effectively use AI tools, individual teams and functions can begin to implement many small but impactful improvements, that can add up to significant benefits.
Where there is a high availability of data, there is the potential to leverage AI and machine learning tools to great effect. Couple this with a high cost, operationally complex industry, and you create significant opportunity to reduce costs, improve end to end processes and enhance the customer experience.
The energy sector is fundamentally different to Financial Services, but there are parallels in the opportunity that AI affords. We find it useful to look across industries to inspire new ideas and identify new opportunities that we can introduce to clients.