The Challenge

RightShip’s Ship Vetting Information System (SVIS™) produced risk ratings based on rules derived from historical ship data combined with expert knowledge and experience. The new RightShip Qi model implements a data-driven solution with machine learning techniques. The new risk rating predicts the propensity for vessel casualty using a complex model calculated from a mastered “golden record” of historical data available on individual vessels from a wide variety of sources. The challenge in rolling out RightShip Qi was to ensure the accuracy and stability of the predictive results while also preserving customer confidence. RightShip needed to visually communicate the risk scoring associated with a vessel in an intuitive fashion to assure their clients that the factors influencing a rating were logical and, where possible, actionable.

The Solution

Elder Research performed a thorough assessment of RightShip’s end-to-end process to measure its compliance with predictive analytics best practices and standards, covering:

  • Data Ingestion
  • Data Warehousing
  • Extract, Load, and Transform (ETL)
  • Model Building
  • Results Presentation
  • Model Updating

Elder Research found that RightShip Qi represented a significant technological improvement over the SVIS baseline. Elder Research also worked with RightShip stakeholders to understand their visualization needs and develop an intuitive presentation of the factors associated with a given risk score as shown in the example below.

RightShip Case Studay Figure 1


Thanks to the thorough Model Validation performed by Elder Research, RightShip was able to release the Qi model with confidence. RightShip also successfully implemented and deployed the developed visualization methodology, providing charterers and vessel owners with better insights on the factors contributing to assigned star ratings for each vessel.

  Download This Case Study