In the decision-making process regarding the allocation of public and private resources, it is crucial for public decision-makers and private operators of the containerized maritime freight market, the knowledge of a good estimation of the demand variables: throughput, transhipment and origin-destination flow.

There is a large amount of works on freight demand models at international scale, which range from aggregated models which provide general estimates at macro-geographical areas and years, to disaggregate models which simulate the individual firm behaviour. Unless the continuous efforts, the nature and intensity of relationships between the throughput variables and the origin-destination flows ones are not sufficiently well established in the existing scientific literature.

On the other hand, international consultants and research institutes periodically provide general estimates. However, the lack of disaggregated data, not easy to obtain from the actors of the market who prefer their activities not to be shared, does not allow estimations to be obtained for specific geographical areas and periods of interest.

The two main objectives of the paper are:

(i) ex-post analysis for historical estimation at a disaggregate level of the demand variables of the containerized maritime freight market;

(ii) ex-ante analysis for forecasting purposes, through the specification and calibration of relationships between the throughput and origin-destination flow variables.

The objectives are achieved by defining a procedure that allows performing spatial and temporal fitting operations from heterogeneous data sources, in order to obtain good estimations of the demand variables referred to specific geographical area and period of interest.

The procedure is applied to obtain estimates of demand variables at 2006 and 2015 for the Mediterranean area, according to expectations before the crisis and then to current developments in the container market.