The power of data: Machine learning and advanced analytics

At DCC, the Demand Management team play an important role in “sizing up” the network, to enable the infrastructure to cope with the rising capacity and demand of Service Requests in the future. Previously, the team would have to engage the customers, who would then provide a forecast of their Service Request volumes over the next eight months. It was very much a manual task, whereby DCC had to contact 50+ customers individually. It was also hugely time consuming for energy retailers, who could have been concentrating on projects of more value to their customers.

The value of DCC’s data cannot be overstated. We exist as a monopoly to the energy industry, and because of this, we are in a unique position with access to data that is brand new, and previously never been available in the UK. The data itself is complicated and unprecedented, and the data volumes we are working with are well within the definition of “Big Data” – we are crunching somewhere around 40 million records of data per day. It is crucial to have skilled Data Analysts, Data Engineers and Data Scientists working with this data, driving insightful analysis and building intelligent data solutions. One of these solutions is the Service Request Forecasting suite.

Service Request Forecasting is a collection of machine learning models, designed to automate the process of sizing up and future-proofing the network. A time series model is created for every energy supplier and every service request, and the outputs are aggregated to give a full network view or analysed at individual supplier or service request level. DS&A are in an exclusive position to produce these outputs; we can utilise patterns from other suppliers and service requests and feed these inputs into each model, increasing the overall intelligence of the AI system.

The Service Request Forecasting suite has been a big win for DCC and our customers. We have removed the human element from forecasting, in that we no longer need to physically ask customers for their forecast files; we now generate these automatically, in house, using historic data stored in our data warehouse. The output files are produced in a specific format and automatically facilitate a number of reports, helping us to meet our Regulatory requirements.

DCC and specifically the Data Science & Analytics team are at the beginning of a journey. The team will strive to add value for our customers by deploying state-of-the-art machine learning and advanced analytics, increasing business accuracy for our customers, which in turn will help homes and small businesses save money, and enable energy suppliers to design products that will improve people's lives.