Digital Analytics in Oil & Gas

Artificial Intelligence and machine learning shed light on lower-level risk

While significant incidents were already thoroughly investigated, this oil and gas company was interested in understanding minor incidents better. What was causing people and environmental control failures? Complex data existed, but there was no way to map against the organization's risk framework to run trend analyses or predict near-future events.

To improve the company's ability to uncover direct and root causes of any incident, dss⁺ was asked to help glean insights from their data, such as design and operational control weaknesses and trends in unsafe acts and conditions.

dss⁺ Approach

First, the dss⁺ team reviewed HSE data quality and usability, noting anomalies or gaps that might get in the way of recognising trends, patterns, relationships and other actionable insights. Key indicators were then drawn from the dataset, such as total and top-five hazard condition types, to populate a list of hazards and associated controls.

Data was then used to generate visualisations, such as keyword clouds, cause-and-control displays, bowtie reports and control-type trends, in response to specific queries. AI modelling, natural language processing and machine learning were also employed, exposing previously untrained entities to further attention. All information is now readily accessible via a client dashboard, driving better and faster decisions.