Ship data
Having comprehensive ship data, including ship dimensions, machinery specifications, and other technical details, is crucial for accurate modeling and assessing ship emission. However, obtaining this information for the entire fleet is often challenging, as it can be scattered across multiple sources, inconsistently reported, or restricted by proprietary databases.
For the MariTeam model, we have gathered extensive ship data from Sea-web™ Ships, a comprehensive maritime database, and compiled it to create a detailed representation of the global fleet. This dataset includes key vessel attributes, allowing for a more complete understanding of fleet composition and capabilities. Our ship database consists of:
Ship types
We have a total of 13 distinct ship types, each representing a key category within the maritime industry. These ship types vary in terms of their design, function, and operational characteristics:
| Ship type | Definition |
|---|---|
| Bulk Carriers | Transport unpackaged bulk cargo like coal, grain, and ore in large cargo holds |
| Container Ships | Carry standardized containers for efficient transport of goods across global trade routes |
| General Cargo Ships | Carry various unpackaged goods, machinery, and bulk cargo often loaded individually or on pallets |
| Refrigerated Cargo Ships | Transport perishable goods like fruits and meat in refrigerated holds or containers |
| Chemical Tankers | Designed for transporting liquid chemicals, often with specialized coatings and segregated tanks |
| Gas Tankers | Transport liquefied gases like LNG (liquefied natural gas) or LPG (liquefied petroleum gas) in pressurize or refrigerated tanks |
| Ro-Ro Cargo Ships | Roll-on/roll-off vessels designed for wheeled cargo like cars, trucks, and trailers, loaded via ramps |
| Cruise Ships | Passenger vessels designed for leisure travel, featuring accommodations and amenities |
| Passenger Ships | Transport people on scheduled services or ferries, often with limited cargo and vehicle capacity |
| Crude Oil Tankers | Transport unrefined crude oil in large quantities from production sites to refineries |
| Oil Product Tankers | Carry refined petroleum products such as gasoline, diesel, and jet fuel |
| Offshore Supply Vessels | Support offshore oil and gas operations by transporting supplies, equipment, and personnel |
| Fishing Vessels | Catch and process fish or seafood at sea, ranging from small boats to large factory trawlers |
Below, you will find a plot that visualizes the different levels at which ships are categorized according to the StatCode 5 Shiptype Coding System (a categorization of ships by type). The plot includes all five levels of the StatCode 5, ranging from Level 1 to Level 5, offering a hierarchical view of the ship types. Users can interact with the plot by clicking on the ship types to expand and reveal more detailed information, or click on them again to collapse and go back to a higher level of the hierarchy. The numbers indicate the number of ships present in our database (including decomissioned ships).
Building history
We can reconstruct the historical trend of ship stock and shipbuilding activity by analyzing when ships were built. The plot below illustrates the number of new builds per year, starting from 1950. Notably, significant events, such as the shipbuilding boom followed by its collapse after the 2008 financial crisis, can be clearly seen in the data. This data offers valuable insights into both market dynamics and the evolution of global shipping capacity and the long-term shifts in the global shipping industry. You can select whether to view the number of ships or total deadweight, which offers a more representative measure of the fleet’s total cargo-carrying capacity.
Ship parameters
Each ship in our database constains 92 different parameters, providing a comprehensive profile description of each vessel’s dimensions, cargo capacity, installed machinery, and operational details. As an example, we display some of this data for three distinct ships: Amber 6, a crude oil tanker; Ore Brasil, a bulk carrier; and MSC Irina, a container ship in the table below.
For more detailed information regarding the meaning of some of these parameters, you can refer to the dropdown menu below.
Deriving missing values
This section is based on the work of Kim et al. (2024) “A novel method for estimating missing values in ship principal data” that made of this ship data to develop a novel model to complete missing ship data. This method has been since then incorporated into our model to fill the gaps of parameters the have faulty data or were originally missing. It consists of three steps:
-
Initial Computation: The goal is to create a complete dataset by filling in missing values using curve fitting techniques. Various function forms (linear, quadratic, etc.) are applied to ship design parameters, and the best-fitting function is chosen based on the coefficient of determination. Missing values are filled iteratively by using the fitted values from the most relevant variables available.
-
Final Imputation: In this step, missing values are updated by performing regression analysis on the completed data from step 1. Variables are adjusted based on their relationships with the target variable using curve fitting. A multiple regression analysis with backward elimination is used to create predictive models for each variable. The best model is selected based on statistical significance, and missing values are replaced with predictions from the model.
-
Minor Adjustment: This step involves checking and correcting any implausible values based on known data ranges or domain knowledge. Invalid values are either re-estimated using domain expertise or replaced with values predicted from earlier steps.
Some of the results are show below:
Author: Diogo Kramel
Model: Ship Data Completer
Repository: GitHub
Data Version: v2.0.0 | 2025-02-13
Latest Update: July 23, 2025
Contact: diogo.kramel@ntnu.no