Vessel portfolio analysis

The marine insurance industry is worth over £33 billion globally, with the Lloyds market taking a prioritised interest in it. There are thousands of data points and sources available to provide rich analysis, should they be utilised in the right way.

I held a workshop with an underwriting assistant to understand what their needs are for the project, revealing the following:

Underwriters insure various fleets of vessels that make up their portfolio. They have access to tools that allow them to view information about an individual vessel, but these tools do not extend to showing how this one vessel affects the risk profile of a wider portfolio of vessels.

Where analysis on a portfolio level is not currently available, my team were mandated to explore a solution.

• Ships in the same zone that may be liable to the same catastrophes

• Ships that frequently take similar routes that may be affected by catastrophes or geopolitical issues that affect their journey.

• Underwriters often may make occasional exceptions, taking on ships that may not meet certain quotas. Though these are sometimes necessary, the underwriter must keep track of the overall balance of their portfolio and ensure that not so many exceptions are being made that the overall risk of the portfolio is enhanced.

• Underwriters must rely on goodwill to be notified of any breaches of their policy where vessels sail through hazard zones - pre-identified areas that have been marked as potential hazards for war or adverse weather.

A dashboard that allows users to assess their insured vessels in the context of their entire portfolio through three different views:

All vessels from a user's portfolio are plotted on a map using live location data accurate up to the last 15 minutes. Users are able to lay active hazard locations over the map of their plotted vessels, enabling them to see where vessels have passed through and check for policy breaches. For example, a user can clearly see and is notified when a vessel passes through a war zone.

• Insured information

  All information about how much risk is covered, including risk score, financial implications

• Vessel particulars

  All technical information about the vessel. This gives the underwriter a good understanding of the type of vessel it is and will help them make decisions on how to insure it.

• Route

  The vessel's route from the last three months, including all of the port calls and lengths of mooring.

• Correlation insights

  Using available data from the IHS, I worked with data scientists to devise AI-driven insights about the level of correlation and similarity within the portfolio. Insurers aim to keep a high diversity in their portfolios to prevent the likelihood of multiple claims occuring from a singular event.

• Comparison

  Users are able to see where this vessel falls compared to the average of both their portfolio and all registered vessels. Using data visualisation, users can easily see at a glance when a vessel falls significantly out of the average.

• Activity log

  A list of recent events affecting the vessel, colour coded to imply their level of importance. Users can order these events by date or by level of severity.

• Implied risk

  Another data science model that predicts the likelihood of different risks and applies a score to individual factors

• Insured statistics

  Users are quickly able to digest the statistics at the top of the page to understand the size of their portfolio and easily track their loss ratio.

• Insured assets

  All vessels listed, with defining financial information extracted for an overview

• Vessel map

  All vessels plotted in their current location on a map, with vessels with recent policy breaches highlighted

• Correlation analysis

  A presentation of the findings of a custom data science model to highlight areas of the portfolio that are lacking in diversity

• Claims history

  The most recent and largest claims coming from the portfolio. Users are able to recognise patterns or increased liabilities that come with each vessel

• Analysis sandbox

  Users are able to perform in-depth analysis on their portfolio and test scenarios, with their financial statistics reflecting the changes. Users can hide vessels from their portfolios, temporarily add in prospective vessels, and edit the premiums, discounts and coverage for vessels, to see how their overall portfolio is affected.

The quality of internal data regarding insured assets is a recurring painpoint. Though the tool has many effective elements that rely on public data, a lot of the value-adding features of the tool rely on internal data. Information such as claims history and insured totals is scattered and varies between different teams. Many of the features of this tool require the users to have a historically consistent approach to storing information in order to be able to cross analyse between different insured assets. In order to ensure these features would go live, a lot of manual work had to cleanse the data to be used efficiently.

Users are easily able to analyse vessels on both a singular and a portfolio level. Where previously, users had to cross reference multiple spreadsheets of their personal portfolio with online tools, they are now able to use a dashboard that utilises AI to present smart insights on their portfolio. Insurers are able to keep track of, analyse and balance out their portfolios easier and faster than ever by using this smart and engaging platform.