how-things-works

The Marine Benchmark Methodology

A structured approach to qualifying, interpreting and developing maritime data into decision-ready intelligence.

A Controlled Framework for Maritime Intelligence

A common methodology governs how data enters, develops within and is used across the Marine Benchmark Data Foundation.

Marine Benchmark's methodology covers the full analytical lifecycle of maritime information — from the qualification of source data to contextual interpretation, derived intelligence, prediction and enterprise use.

The sections below describe the principles that govern this process, including how information is maintained through time, connected across the platform, strengthened through research and domain knowledge, and kept traceable to its underlying maritime evidence.

From Source Data to Maritime Intelligence

Source data enters Marine Benchmark only after it has been qualified, normalised and placed in maritime context.

Maritime source data varies substantially in coverage, frequency, quality, definition and technical format. Marine Benchmark therefore treats each incoming source stream as information to be validated, reconciled, normalised and contextualised before it enters the analytical process.

The challenge extends well beyond AIS. Maritime datasets range from high-frequency vessel observations and changing reference data to large environmental datasets distributed in specialised scientific and binary formats. Bringing these sources into a common analytical environment requires sustained processing, quality control and maritime interpretation.

Terrestrial and satellite AIS, vessel information, weather and ocean data and other maritime sources are assessed against maritime domain knowledge built through teams with decades of industry experience. This provides a controlled basis for identifying inconsistencies, resolving meaning and determining how different sources can be used together.

Only after this qualification does source data become part of the Data Foundation and contribute to derived maritime intelligence. Observed, derived and predictive information remain distinct throughout the platform, preserving the evidential basis of each.

A Structured Model of the Maritime World

Maritime activity becomes more useful when it can be related to the physical, commercial and regulatory environment in which it occurs.

Maritime activity becomes more useful when it can be related to the physical, commercial and regulatory environment in which it occurs.

Marine Benchmark maintains approximately 70,000 maritime geographic objects, including around 9,000 ports, 17,000 terminals and 41,000 named berth polygons, together with anchorages, yards, passages, canals and straits, offshore assets, regulatory areas, war and risk zones and other maritime geographies.

This geographic knowledge is connected with vessel, stakeholder, regulatory and other domain information used across the platform.

The resulting structure supports analysis from global shipping and fleets to individual vessels, voyages, ports, terminals, berths and events.

Time-Aware by Design

Marine Benchmark maintains global maritime history back to 2009.

The maritime system changes continuously. Fleets evolve, vessel and stakeholder relationships change, infrastructure develops, and geographic, regulatory and risk environments change over time.

Where historical context is relevant, Marine Benchmark maintains information through time rather than assuming that the current state of the maritime world also describes the past.

Historical and current information can therefore be analysed within a common framework while retaining the context applicable to the period being studied.

History Remains Live

Historical information is part of the active Marine Benchmark Data Foundation rather than a separate archive.

The full historical environment remains hot and instantly accessible, regardless of whether the underlying activity occurred today or more than a decade ago.

Marine Benchmark's analytical environment contains tens of billions of rows across hundreds of data fields, maintained for instant access and analysis. The Weather & Ocean data alone comprises approximately tens to hundreds of billion rows, all maintained as hot data.

This allows long historical series to be accessed alongside current information without moving between separate current and historical data environments.

The processing architecture also allows historical information to be reprocessed as methodologies and underlying information develop. Relevant improvements can therefore be reflected historically rather than creating unnecessary discontinuities between older and newer data.

For organisations using maritime data in quantitative research, enterprise applications, models and AI, this provides a consistent analytical environment across time.

Vessel-Specific. Globally Consistent.

Marine Benchmark applies common definitions across the global fleet while accounting for the characteristics and circumstances of individual vessels.

This principle extends across vessel activity, voyages, capacity, performance, energy consumption, emissions and other derived maritime information.

It provides a common analytical basis for comparing vessels, fleets, markets, geographies and periods while retaining the granularity of the underlying maritime activity.

Information can be analysed at different levels of the maritime system and investigated towards the vessels, voyages and events contributing to it.

Knowledge Compounds

Marine Benchmark is built as a connected Data Foundation rather than a collection of independently processed datasets. Information developed in one part of the platform can therefore strengthen interpretation and analytical output elsewhere.

These relationships are not applied generically. They are governed by maritime domain knowledge, which determines which connections are economically, operationally and technically meaningful and when different types of information should be considered together.

AI has become an increasingly important tool for identifying, testing and establishing relationships across the platform. As the number and complexity of potential relationships increase, maritime domain knowledge becomes more important in validating relevance and distinguishing meaningful structure from statistical coincidence.

As new validated relationships are incorporated, the knowledge embedded within the Data Foundation compounds. Each additional connection increases the context available elsewhere in the platform, strengthening the analytical value of the existing data and creating a progressively richer representation of the maritime system.

A Continuous View Through Time

Marine Benchmark's Data Foundation follows maritime activity continuously through time — from historical observations and reconstructed activity, through the current state of a vessel, to its expected movement and next destination.

More than 17 years of history form the analytical base. Current positions, events and voyages extend that history into the present, while predictive intelligence carries the same vessel context forward including expected destination, arrival time, future position and anticipated passages or events along the route.

This creates a continuous operational timeline in which history, now and what happens next are connected rather than treated as separate datasets.

Longer-term outlooks are maintained separately. Fleet Evolution & Forecast addresses structural change over years and decades, including future vessels, capacity, technology, renewal and retirement.

Traceability and Method Development

Marine Benchmark's platform and methodologies have been developed over approximately 15 years through commercial application, applied research, academic collaboration and maritime analysis.

Academic research strengthens our methodologies, while Marine Benchmark provides the ability to translate research findings and regulatory changes into fleet-wide, regional and global impact analysis. This allows academic results to be tested against the physical characteristics and activity of the world fleet.

One example is the development of load-dependent NOx emission factors, informed by approximately 30,000 sniffer measurements from university research and incorporated into vessel-level modelling across the Data Foundation.

The same analytical foundation has supported work involving organisations including the International Maritime Organization and European Commission, connecting empirical research, maritime modelling and regulatory impact analysis.

Derived intelligence remains connected to the underlying maritime information from which it was constructed, supporting investigation and verification from global results down to the contributing vessels and activity.

One Data Foundation. Nine Data Domains.

Marine Benchmark organises its customer-facing data into nine connected Data Domains.

Fleet Structure & Capacity

Fleet composition, characteristics and physical capacity.

Fleet Evolution & Forecast

Fleet renewal, retirement, technology and long-term development.

Positions & Events

Historical and current positions, movement, vessel events and forward-looking movement intelligence.

Voyages & Vessel Activity

Operational activity and reconstructed commercial voyages.

Cargo & Trade

Physical cargo and commodity movements.

Capacity & Performance

Deployment, utilisation, transport work, congestion and performance.

Weather & Ocean

Environmental conditions affecting vessels and maritime assets.

Energy, Emissions & Regulation

Energy use, fuel consumption, emissions and regulatory exposure.

Risk & Compliance

Maritime risk, geographic exposure, stakeholder context and compliance intelligence.

Designed for Enterprise Use

Marine Benchmark is built as a common data environment for applications, analytical workflows and enterprise systems.

Our own applications use the same APIs and underlying Data Foundation available to customers. The performance required to load complex analytical views in Marine Benchmark applications is therefore available for integration into customer systems as well. A dashboard can, for example, retrieve multiple long historical time series within seconds.

Customers can combine access methods according to workflow: bulk or analytical downloads for research and modelling, API access for ad hoc queries and internal dashboards, cloud and data feeds for systematic integration, and Marine Benchmark applications for direct exploration, verification and validation.

The same underlying intelligence is available across these delivery modes. Data downloaded into an internal analytical environment can therefore be examined against the corresponding vessels, voyages, events and market context in Marine Benchmark applications without moving between separate data products.

Discuss the Data Foundation with us

For technical, methodological or analytical questions, speak directly with the team behind Marine Benchmark.

If you are evaluating Marine Benchmark for enterprise integration, quantitative research, model development or specialist maritime analysis, we can go deeper into the areas relevant to you — from data coverage and methodology to traceability, performance, access and implementation.