The Gregario analytics team is part of DataTalk, a provider of business information analytics to develop and improve communications with customers, prospects and employees. It brings statistical techniques to bear on data to identify key features and actionable insights.

The largest private global shipping broker of bulk cargo commissioned DataTalk to apply data science to the prediction of global bulk shipping movements for 784 supertankers across a six-month period. Carrying crude or refined oil, the daily movement of these vessels has an opportunity cost of £500,000.

 

Using location data which is manually transmitted by each ship every two hours, triangulation allows for latitude and longitude to be calculated, with depth, bearing, destination and expected arrival time added manually. With some $400 million-worth of cargo on-board, any unexpected changes in direction can be important for markets and future trades.

 

An agent-based modelling approach for each vessel was applied to allow for anticipation of market trends as well as geo-political events. This allowed for the detection of flows of oil from the USA as a result of shale gas exploration as well as the return of deep drilling in the Gulf to be identified. The project proved a potential $1 billion improvement in market efficiency from the use of arrival time predictions from data science, rather than manual calculations.

Oil is the new data! A major proof-of-concept that could have a huge pay-off. Impressive work.