These materials are provided by the Georgia Tech Panama Logistics Innovation & Research Center. You are welcome to use them so long as the copyrights remain intact and credit for authorship is acknowledged. Read the full paper here.

The Global Container-Shipping Network

The movement of shipping containers traces out the paths of goods movement. This reflects levels of trade and, more generally, economic standing. Here we provide a display of the container-shipping network that moves goods. It shows most of the direct container-shipping links between all pairs of the world's almost 400 container ports. The display is via a computer program that you can run from here.

The Georgia Tech “Container Port Connectivity Index”

Besides displaying the network, the program reports statistics about the container capacity of shipping along each link and also statistics about each port. For each link one can read the median transit time, the average weekly capacity measured in TEUs, and the average weekly number of reefer plugs available.

This information and more is combined for each link to create a “weight” that reflects the intensity of flow of container capacity along that link. (The computation follows that of the Liner Shipping Connectivity Index suggested by United Nations Conference on Trade and Development (UNCTAD) to model trade between nations.) Then we use these weights to compute a score for each port that reflects the strength of its connectivity to other ports within the network. We call this score “Container Port Connectivity Index” (CPCI). In fact, the Container Port Connectivity Index consists of two numbers, one that reflects how well connected the port is for imports and the other for exports.

What is distinctive about our index is that it is based on more than just local data. The computation depends not just on the immediate trading neighbors of a port, but also on the scores of those immediate neighbors, and the scores of the neighbors of the neighbors, and so on.

You can find all the technical details here.

View the global container-shipping network

We provide a computer program that displays the global container-shipping network and allows you to interact with it. In addition it evaluates and displays the Port Connectivity Index of all container ports. We will continue to add new features and capabilities to the program. Suggestions are welcomed.

Please read the license and disclaimers, then click to launch the latest version via Java Webstart:

Note 1: This application is unsigned and so may not run unless you give it permission. Under Mac OS X go to System Preferences, choose Security and Privacy and click the button to “Allow apps downloaded from Anywhere”. (Be sure to return this to its original setting when you are done running this application.) For Windows users, follow the instructions here and add the url for the program to the list of trusted url's.

Note 2: Some company firewalls may still prevent Java webstart programs from running, in which case see your system administrator.

If you are running the program for the first time, Java Web Start will download it (5 jar files totaling about 3MB). The next time it will check only for modified jar files (an upgrade) and download them. If there has not been an upgrade, the application will start immediately.

This program is written in Java so it runs on any brand computer and any operating system. If you do not already have Java installed, get the latest version of the Java Runtime Environment (JRE) here:

The display

All significant container ports are plotted by their longitude and latitude. Each port is drawn as a filled disk of size scaled according to a statistic that can be selected by the user.

A link from port A to port B indicates that there is container service directly from A to B. The link is drawn to represent the fact of a direct service — the link does not show the navigational route of a ship. (This is why some links cross over land masses.)

The darker and wider the link, the greater the strength of that link as measured by a calculation mimicking that of the Liner Shipping Connectivity Index. As expected, there are thicker, darker links among the major ports of East Asia and from those ports to major ports in Europe and the US.

How to use the program

Click-and-hold to drag the network; use the mouse wheel or trackpad or menu or specified keystrokes to zoom.
Hold the mouse over a node or link to see summary statistics.
Use the slider at the bottom of the screen to scale ports according to the currently selected statistic.
Click on the radio button on the lower right to change the pointer mode. If it is set to select, you can choose a port by clicking on it and additional ports by shift-clicking. Then you filter the network to display only ports that ship directly to or receive shipments directly from selected ports. You can also use the File menu to save selected ports and later recall them. All reports flag any selected ports so you can identify the relevant results.

Features

The program starts with Europe and Africa in the middle of the screen. From the “View” menu select “Shift display” to toggle between this view and one with East Asia in the middle.
The program computes the shortest travel time from/to any selected port to/from any/all other ports. (Note that distances are not symmetric. Also, the program reports only travel times and does not include time in port, nor does it identify when transshipment is necessary.) For a specific pair of ports, the program will show the sequence of ports that form the quickest route between the two.
The program computes a variety of measures by which the centrality of a port may be judged. Note especially the Container Port Connectivity Index (CPCI), which measures how well integrated a port is into the global network of container-shipping.

Data

The program is powered by data derived from www.compairdata.com and represents the network as it existed in September 2013. (Thanks to Pablo Achurra and his team in Panama for organizing and formatting this data.)

WARNING: We are still refining the database on which these computations are based and so values may change.