Maritime Trends: Why Is AI-Based Predictive Analytics the Next Key Change Driver in the Shipping Industry?

Maritime Trends: Why Is AI-Based Predictive Analytics the Next Key Change Driver in the Shipping Industry?

Dorota Owczarek - June 2, 2022

The maritime industry is currently going through a period of significant change. With new technologies emerging and the container shipping crisis, predictive analytics is becoming an increasingly important part of the shipping process. In this article, we will explore the impact of predictive analytics in the maritime industry and discuss why it is such an important tool for shipping companies.

Container Shipping Industry Shape

Even though today we hardly can imagine a global economy without them, container ships are with us for a relatively short time. The idea of a standardized ship container first emerged and was implemented only in the fifties. It quickly spread outside of America, becoming a foundation of the global maritime supply chain. With their rising popularity, the ships also have been growing in size, enabling increasingly effective global transport.

Today, the largest can fit tens of thousands of containers. However, in recent years, the success story of the maritime industry has taken an unexpected turn. What’s behind the crisis that is already affecting the economy worldwide? We’ll try to break it down, analyzing the most possible scenarios and most promising solutions to that issue.

Container Shipping Crisis and Its Impact on Global Supply Chain

The maritime industry is currently facing a container crisis. The global shipping sector is struggling to keep up with the surge in demand for containerized cargo, and it seems that the only solution is to increase capacity. However, this would require significant investment in new ships and port infrastructure, which may not be feasible in the current market conditions.

At first, the remedy seems almost intuitive. Why don’t the carriers invest in even larger ships? This solution, however, causes new issues to emerge. Once the ship gets too big, it becomes difficult to control it. When the vessel comes across some natural barrier or turbulent weather, a part of the cargo may fall into the water, causing losses counted in millions of dollars.

Our sea routes aren’t prepared for the ships’ further growth in size, either. Meanwhile, the implications of the blockage of the major sea route can seriously put the balance of the global trade in danger. We have seen it happening in March 2021, during the Suez Canal obstruction. Even though this crucial maritime trading route has been blocked for less than a week, it has resulted in record increases in the price of oil and disruption of the supply chains that took months to fix.

The expenses for transporting containers have increased substantially during the previous two years as consumers altered their buying habits to prefer tangible products. Consumers are now spending more of their disposable income on services rather than household items, and interest rates are increasing.

The reduction in usage of shipping containers, owing to global events including Russia’s invasion of Ukraine and China’s lockdowns, has freed up supply. As a consequence, the containers are currently at their lowest rates since June 2021, according to the Freightos worldwide container freight index.

All in all, the price drop has been hampered by bottlenecks and delays on land transport networks. Despite this, costs remain above pre-epidemic levels.

The mixture of high volume, China covid lockdowns, and the war in Ukraine continues to point to challenging global supply chain performance in 2022.

On the other hand, in terms of environmental protection, packing up as much cargo as possible on one cruise seems the best solution. Increasing the number of routes equals more pollution and more carbon dioxide released into the atmosphere, even if the route itself gets optimized.

Another issue is the lack of workforce to take care of the cargo once delivered to the port. It has been getting worse in recent years due to a practice of so-called front-loading. Facing the changing tariffs, the retailers have started to send excessive numbers of containers just to anticipate the higher fees. Once the containers would arrive to the terminal, there was not enough staff to handle them.

What’s the answer for these issues? From our perspective, all hope is in advanced, AI-driven analytics and automation.

Data Analytics in the Maritime Industry – Why Does It Matter?

Due to its complexity and subjectivity to various dynamically changing factors, the maritime industry is a tough nut to crack when it comes to optimization. Long routes with numerous chokepoints, unpredictable weather conditions, and limited cargo offloading options – all these aspects require advanced analytics to prevent delays and financial losses. In order to deal with large and unstructured data sets effectively, employing AI fuelled big data analytics is a must.

Compared to land transport, the sea cargo carriers have much less flexibility in terms of route planning and adjusting to events in real-time. Thus, they should be prepared for different scenarios. With artificial intelligence, they can predict them and test them to minimize losses in the case of downtimes, accidents, or route blockages. The latter rarely happens, but when it does, a big part of the global sea trade can get paralyzed for days.

Data analytics in the maritime industry can solve issues on the side of cargo ships, but also the ports that have been lately struggling with the lack of workforce and effective logistics of the offloading process.

Artificial Intelligence and Business Intelligence – How Deep Learning Models Can Make a Difference?

Forecasting processes are essential to the maritime supply chain. Whereas transport with land trade routes is more flexible in terms of planning due to shorter distances and less cargo, the giant ships usually take weeks to arrive at a final destination with long periods without stops. Such limited ability to react to real-time changes requires much more anticipated and detailed planning based on accurate forecasts. That’s crucial in the times of just-in-time supply chain management, which is very efficient but also much more susceptible to failure. With time-series data, the carriers can observe seasonality and trends to adjust their planning to different factors and avoid issues like route blockages or overcrowded ports.

More commonly, the data scientists pick deep neural networks for regression problems instead of relying on the traditional approach to time series forecasting (univariate models like ARIMA, Prophet, and multivariate ones like Vector Autoregression). That’s because they are not recursive like the traditional time series models. That means they enable us to receive long-term forecasts directly instead of carrying out a series of predictions, each based on the previous outcome.

Another thing that speaks in their favor is the ability to use time-varying variables as features. The neural networks deal better with a mixture of linear and non-linear trends, delivering more precise predictions as a result. Whereas single machine learning algorithms do not outperform traditional time series methods, deep learning has such potential.

Predictive Analytics in Shipping – Industry Use Cases

Considering that all the elements of the supply chain can benefit from advanced predictive analytics, we’ve covered the most common use cases to give you the big picture of its positive impact.

Predictive Analytics in Shipping - Use Cases

Predictive Analytics in Shipping - Use Cases

Predicting Container Supply and Demand

As we already mentioned, the reduced flexibility of maritime transport makes accurate predictions essential. For the container carriers, having access to precise forecasts on supply and demand is a priority as it conditions their financial safety and efficiency. Particularly nowadays, when the global supply chains are getting increasingly affected by the changing geopolitical situation, predictive analytics becomes a powerful weapon against business trouble.

Using socio-economic data (GDP, population, median income, employment rate, etc.) as input and pairing them with detected trends and seasonality in the time series data as well as geopolitical factors, shipping companies estimate the realistic container demand instead of relying on historical data.

For ports and shipping companies, predictive modeling is a win-win. Having accurate predictions, the carriers are able to reduce the number of empty containers and prepare for the increased demand with additional investments or routes. The ports, on the other hand, end up being at lower risk of congestion since the shipping companies feel safe enough to rely on just-in-time management.

Predicting Shipping Locations and Timing

Everyone – the on-land delivery companies, wholesales and e-commerce shops, and, of course, the final customer – wants to know when and where the shipped goods will arrive. A big part of the consumers may even abandon their shopping cart if they do not get the shipping time estimation right away. Introducing predictive analysis for timing and locations is thus in everyone’s interest.

After identifying the relevant inputs (which may include such features as gross weight, route, the destination port, and so on), the model trained for time or location prediction can come up with accurate output that help the carriers plan their routes, and the ports – effectively manage the offloading process. There are quite some machine learning models that can serve that purpose. Random forest and linear regression seem to be the most popular ones with their relatively fast training.

Related case study: Developing a logistics platform offering real-time visibility and integrations with different carriersOne of our clients was seeking to improve the global supply chain optimization productOur challenge? Providing visibility and data transmission for maximum efficiency and control. We supported solution development for end-to-end execution of logistics activities in Supply Chain Management at the PO/SKU level, including PO creation, stock management, suppliers and distributors management, consolidation and load planning, carrier allocation, documentation, and final delivery. Read more about this case study.

Dynamic Pricing Models for Container Shipping

Even though dynamic pricing model isn’t favored by the customers, it’s becoming increasingly common. With so many dynamically changing factors influencing the final cost of the shipping, the dynamic pricing models save the shipment carriers from operational inefficiency. In such competitive times, they cannot just preventively overcharge, and charging too little compromises their financial safety.

How dynamic pricing strategy works? The model estimates how changing variables impact the price and possible demand for profit maximization.

How dynamic pricing strategy works? The model estimates how changing variables impact the price and possible demand for profit maximization.

With dynamic pricing, shipping companies are able to adjust to the demand-supply dynamic. Such models estimate the cost of the shipping service in real-time, updating the prices at least a few times per day. For example, when the cost of oil suddenly peaks – what has happened a few times in recent years – the model adjusts the estimation, preventing the customers from ordering service at an irrelevant price.

Machine learning algorithms get trained with structured and unstructured data (including the time-oriented and location-specific data) to find the correlations between prices and different variables (like route, distance, fuel cost, demand, season, etc.). This way, they learn to predict relevant prices that fuel sales while sustaining financial security and growth.

To dig deeper into these subjects, we recommend our article on dynamic pricing in logistics with recommendations on building a dynamic pricing strategy with artificial intelligence.

Improving Port Operations and Cargo Handling Velocity

The role of artificial intelligence goes beyond vessel-related processes. In recent years, we’ve seen ports struggling to cope with the increasing numbers of arriving containers. The already mentioned front-loading practice (early shipping that prevents low time performance in the case of high demand and avoids higher tariffs) has led to congestion in North American and Asian ports last year. These issues continue occurring as ports lack the equipment and workforce to streamline port operations.

They could, however, be avoided by automation. Some ports in the world – like Tuas mega port in Singapore or Chinese Qingdao – are 100% automated, which reduces the risks of human mistakes that could lead to congestion. Others, like Rotterdam, value the community factor but still use artificial intelligence to optimize their management. Using a digital twin (such tools are not reserved for ships, the land facilities take advantage of them, too), they monitor all their operations in real-time. The digital twin can also serve as a testing ground, enabling ports to check their capacities in different scenarios,

With automated vehicles, the ports can solve the workforce shortage issue and reduce the probability of accidents. AI can also help the container terminals unlock the shipping space for more containers by finding the most efficient configurations.

Predicting Container Availability at Port

Artificial intelligence can also support smoother integration of the sea and on-land logistics companies. The carriers that pick up the slack after the shipping vessels use predictive models to estimate when exactly the container will be ready for a pickup at the port. That facilitates fleet management optimization, helping on-land delivery firms cut costs and distribute their workforce and equipment in a more effective way.

Implementing AI in Shipping

Each project is different, but without a doubt, the shipping industry often requires engaging models that can handle more complex unstructured datasets. As a result, it has adapted deep learning to a much greater extent than the other sectors. Since in this case, the input data is usually gathered via different channels (including sensors collecting information in real-time, maritime weather stations, port systems, or navigation systems),  the developers need to put a strong emphasis on seamless integration.

Without effective processing of data from different sources, of varied qualities and structure, it’s hard to come up with accurate predictions. In this case study, you can read about the related challenges we faced when creating a platform for carriers (including the maritime vessels) and how we approached them.

Many ports in the world, particularly the Asian ones, are heading towards complete automation, at least in the terminal area where the offloading processes take place. That trend will most likely continue since it significantly speeds up container management and reduces the probability of blockages caused, for instance, by vehicle accidents or a shortage of staff.

Many of these automated terminals are prepared to cooperate with autonomous ships. So far, they are a minority, but carriers have been showing a strong interest in employing such a solution in the nearest future. Thus, independent ships may likely follow in the footsteps of autonomous vehicles. It’s worth noting that this category covers entirely autonomous shipping vessels, as well as remotely controlled ones.

With the most recent events (Covid pandemic, Suez Canal congestion), the maritime industry has started to pay more importance to automated crisis management. Until recently, the carriers would focus on productivity, but the instability of the global economy and geopolitics enforce increased investment in risk-detection systems. With increasing probability of climate change-fuelled extreme weather conditions and chokepoint blockage, the carriers will most likely invest heavily in the optimization of their route planning models that employ advanced predictive analytics capable of detecting more subtle correlations and trends.

If you are searching for a way to solve an industry-related problem that has been affecting your company or have innovation in mind, reach out to us. nexocode is a company experienced in developing various predictive analytics solutions and other AI-based systems for logistics and supply chain management. We’d love to hear about it and suggest you some solutions based on our know-how!

About the author

Dorota Owczarek

Dorota Owczarek

AI Product Lead & Design Thinking Facilitator

Linkedin profile Twitter

With over ten years of professional experience in designing and developing software, Dorota is quick to recognize the best ways to serve users and stakeholders by shaping strategies and ensuring their execution by working closely with engineering and design teams.
She acts as a Product Leader, covering the ongoing AI agile development processes and operationalizing AI throughout the business.

Would you like to discuss AI opportunities in your business?

Let us know and Dorota will arrange a call with our experts.

Dorota Owczarek
Dorota Owczarek
AI Product Lead

Thanks for the message!

We'll do our best to get back to you
as soon as possible.

This article is a part of

AI in Logistics
51 articles

AI in Logistics

Artificial Intelligence is becoming an essential element of Logistics and Supply Chain Management, where it offers many benefits to companies willing to adopt emerging technologies. AI can change how companies operate by providing applications that streamline planning, procurement, manufacturing, warehousing, distribution, transportation, and sales.

Follow our article series to find out the applications of AI in logistics and how this tech benefits the whole supply chain operations.

check it out


Insights on practical AI applications just one click away

Sign up for our newsletter and don't miss out on the latest insights, trends and innovations from this sector.


Thanks for joining the newsletter

Check your inbox for the confirmation email & enjoy the read!

This site uses cookies for analytical purposes.

Accept Privacy Policy

In the interests of your safety and to implement the principle of lawful, reliable and transparent processing of your personal data when using our services, we developed this document called the Privacy Policy. This document regulates the processing and protection of Users’ personal data in connection with their use of the Website and has been prepared by Nexocode.

To ensure the protection of Users' personal data, Nexocode applies appropriate organizational and technical solutions to prevent privacy breaches. Nexocode implements measures to ensure security at the level which ensures compliance with applicable Polish and European laws such as:

  1. Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation) (published in the Official Journal of the European Union L 119, p 1); Act of 10 May 2018 on personal data protection (published in the Journal of Laws of 2018, item 1000);
  2. Act of 18 July 2002 on providing services by electronic means;
  3. Telecommunications Law of 16 July 2004.

The Website is secured by the SSL protocol, which provides secure data transmission on the Internet.

1. Definitions

  1. User – a person that uses the Website, i.e. a natural person with full legal capacity, a legal person, or an organizational unit which is not a legal person to which specific provisions grant legal capacity.
  2. Nexocode – NEXOCODE sp. z o.o. with its registered office in Kraków, ul. Wadowicka 7, 30-347 Kraków, entered into the Register of Entrepreneurs of the National Court Register kept by the District Court for Kraków-Śródmieście in Kraków, 11th Commercial Department of the National Court Register, under the KRS number: 0000686992, NIP: 6762533324.
  3. Website – website run by Nexocode, at the URL: whose content is available to authorized persons.
  4. Cookies – small files saved by the server on the User's computer, which the server can read when when the website is accessed from the computer.
  5. SSL protocol – a special standard for transmitting data on the Internet which unlike ordinary methods of data transmission encrypts data transmission.
  6. System log – the information that the User's computer transmits to the server which may contain various data (e.g. the user’s IP number), allowing to determine the approximate location where the connection came from.
  7. IP address – individual number which is usually assigned to every computer connected to the Internet. The IP number can be permanently associated with the computer (static) or assigned to a given connection (dynamic).
  8. GDPR – Regulation 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of individuals regarding the processing of personal data and onthe free transmission of such data, repealing Directive 95/46 / EC (General Data Protection Regulation).
  9. Personal data – information about an identified or identifiable natural person ("data subject"). An identifiable natural person is a person who can be directly or indirectly identified, in particular on the basis of identifiers such as name, identification number, location data, online identifiers or one or more specific factors determining the physical, physiological, genetic, mental, economic, cultural or social identity of a natural person.
  10. Processing – any operations performed on personal data, such as collecting, recording, storing, developing, modifying, sharing, and deleting, especially when performed in IT systems.

2. Cookies

The Website is secured by the SSL protocol, which provides secure data transmission on the Internet. The Website, in accordance with art. 173 of the Telecommunications Act of 16 July 2004 of the Republic of Poland, uses Cookies, i.e. data, in particular text files, stored on the User's end device.
Cookies are used to:

  1. improve user experience and facilitate navigation on the site;
  2. help to identify returning Users who access the website using the device on which Cookies were saved;
  3. creating statistics which help to understand how the Users use websites, which allows to improve their structure and content;
  4. adjusting the content of the Website pages to specific User’s preferences and optimizing the websites website experience to the each User's individual needs.

Cookies usually contain the name of the website from which they originate, their storage time on the end device and a unique number. On our Website, we use the following types of Cookies:

  • "Session" – cookie files stored on the User's end device until the Uses logs out, leaves the website or turns off the web browser;
  • "Persistent" – cookie files stored on the User's end device for the time specified in the Cookie file parameters or until they are deleted by the User;
  • "Performance" – cookies used specifically for gathering data on how visitors use a website to measure the performance of a website;
  • "Strictly necessary" – essential for browsing the website and using its features, such as accessing secure areas of the site;
  • "Functional" – cookies enabling remembering the settings selected by the User and personalizing the User interface;
  • "First-party" – cookies stored by the Website;
  • "Third-party" – cookies derived from a website other than the Website;
  • "Facebook cookies" – You should read Facebook cookies policy:
  • "Other Google cookies" – Refer to Google cookie policy:

3. How System Logs work on the Website

User's activity on the Website, including the User’s Personal Data, is recorded in System Logs. The information collected in the Logs is processed primarily for purposes related to the provision of services, i.e. for the purposes of:

  • analytics – to improve the quality of services provided by us as part of the Website and adapt its functionalities to the needs of the Users. The legal basis for processing in this case is the legitimate interest of Nexocode consisting in analyzing Users' activities and their preferences;
  • fraud detection, identification and countering threats to stability and correct operation of the Website.

4. Cookie mechanism on the Website

Our site uses basic cookies that facilitate the use of its resources. Cookies contain useful information and are stored on the User's computer – our server can read them when connecting to this computer again. Most web browsers allow cookies to be stored on the User's end device by default. Each User can change their Cookie settings in the web browser settings menu: Google ChromeOpen the menu (click the three-dot icon in the upper right corner), Settings > Advanced. In the "Privacy and security" section, click the Content Settings button. In the "Cookies and site date" section you can change the following Cookie settings:

  • Deleting cookies,
  • Blocking cookies by default,
  • Default permission for cookies,
  • Saving Cookies and website data by default and clearing them when the browser is closed,
  • Specifying exceptions for Cookies for specific websites or domains

Internet Explorer 6.0 and 7.0
From the browser menu (upper right corner): Tools > Internet Options > Privacy, click the Sites button. Use the slider to set the desired level, confirm the change with the OK button.

Mozilla Firefox
browser menu: Tools > Options > Privacy and security. Activate the “Custom” field. From there, you can check a relevant field to decide whether or not to accept cookies.

Open the browser’s settings menu: Go to the Advanced section > Site Settings > Cookies and site data. From there, adjust the setting: Allow sites to save and read cookie data

In the Safari drop-down menu, select Preferences and click the Security icon.From there, select the desired security level in the "Accept cookies" area.

Disabling Cookies in your browser does not deprive you of access to the resources of the Website. Web browsers, by default, allow storing Cookies on the User's end device. Website Users can freely adjust cookie settings. The web browser allows you to delete cookies. It is also possible to automatically block cookies. Detailed information on this subject is provided in the help or documentation of the specific web browser used by the User. The User can decide not to receive Cookies by changing browser settings. However, disabling Cookies necessary for authentication, security or remembering User preferences may impact user experience, or even make the Website unusable.

5. Additional information

External links may be placed on the Website enabling Users to directly reach other website. Also, while using the Website, cookies may also be placed on the User’s device from other entities, in particular from third parties such as Google, in order to enable the use the functionalities of the Website integrated with these third parties. Each of such providers sets out the rules for the use of cookies in their privacy policy, so for security reasons we recommend that you read the privacy policy document before using these pages. We reserve the right to change this privacy policy at any time by publishing an updated version on our Website. After making the change, the privacy policy will be published on the page with a new date. For more information on the conditions of providing services, in particular the rules of using the Website, contracting, as well as the conditions of accessing content and using the Website, please refer to the the Website’s Terms and Conditions.

Nexocode Team