Logistics & Supply Chain

Empowering a Shipping Provider to Leverage AI to Improve Accuracy of Demand Insights - AI Use Case in Supply Chain

PI.EXCHANGE worked with a shipping company to present a solution to an AI use case that predicts the itinerary of competitors' vessels and time at the port

Attaining and maintaining a competitive edge in the logistics industry is dependent on building efficiency into operations. This includes an accurate understanding of demand. Part of understanding demand is accurate and timely information on competing providers’ vessels, their location and time in port.

Working with an Intermodal maritime logistics company (the Client), PI.EXCHANGE presented a resource-efficient solution to predict the itinerary of competitors' vessels and time at port. This enabled a better understanding of the short-term fluctuations in global vessel movements so that the client could better predict demand and coordinate their fleet for maximum asset and capacity utilization with minimal idle time.

For other relevant use cases, check out our article on the Top 18 essential AI Use Cases in Leading Industries!


The objective: The Client wants to use predictive modelling to improve the speed and accuracy of forecasting demand to position and coordinate their fleet of vessels efficiently and confidently.


The logistics industry has access to an increasing volume of data, their challenge is that the conventional methods to extract business intelligence have required large investments in internal resources, expertise and infrastructure in order to synthesize all of the data and enable data-driven decision-making.

The global value of goods transported according to HSBC’s Trade Flow Report (2017) is expected to develop to a value of US$68.5 trillion by 2050.

The Client provided historical satellite data to predict the itinerary of competitors, so decision-makers could be best equipped to provide assets where needed, when needed. 


One year of historical shipping data was provided for 10 vessels. The data included all the sightings (provided by satellite data) and associated attributes like longitude and latitude, estimated time of arrival, navigation status distance to the port, and region of the port.

The shipping data was ingested cleaned, merged and transformed using the smart data preparation feature.  During this process, 10 disparate datasets become 1 dataset,  and using existing data, new features were constructed to assist with the modelling process. These included; next port and time spent at each port.

Once the data set was prepared, 5 models were created using the model recommender to select and train the optimal models given the datasets. These models were to:

  • Predict the next port for a vessel;

  • Predict the 2nd port for a vessel;

  • Predict the 3rd port for a vessel;

  • Predict the 4th port for a vessel; and

  • Predict the time spent at the port, (used in conjunction with the other models).

Models were then deployed to the cloud for testing and validation on out-of-time data. The insights were then used to augment human decision-making with predictive intelligence. Essentially, using PI.EXCHANGE’s AI & Analytics Engine, the Client could quickly model historical shipping data to understand competing vessels' itinerary and coordinate their own vessels, according to the predicted demand.

“The platform offered numerous machine learning models. We were offered to trial our historical data sets with each model to identify which of them offered the most accurate forecast. 

Chartering Manager

AI & Analytics Engine model deployment process


PI.EXCHANGE’s AI & Analytics Engine provided the Client with: 

  • 90% accuracy of predicting the probability of next port for 4 ports (one-month itinerary.) 
  • 85% accuracy at predicting the days spent at a port by competitors vessels.

Additional accomplishments include:

  • Minutes vs Weeks: Reduced data analysis and model building process cycle from several weeks to minutes. 
  • Fast and accurate predictions, compared to reliance on domain knowledge alone.
  • Intuitive collaboration and management: Enabled a single dashboard view for improved and simplified ML development and collaboration, improving data team output.

For a related Case Study on the use of AI in Logistics, check out this article!


The Client was able to see demonstrable value in augmenting their internal capabilities with a layer of AI-powered intelligence. By understanding where their competitors were going and for how long, in a timely and accurate manner,  The client was able to save significant resources by confidently optimizing their fleet positioning, to decrease the amount of idle time and improve asset utilization. 

The models, once deployed, are automatically monitored for drift over time. This provided the Client with a clear picture of model health for continued optimal performance from ML projects.

Machine Learning Model's health metrics

The platform demonstrated value in complementing my team’s decision-making process.”

Chartering Manager

The value of the PI.EXCHANGE’s AI & Analytics Engine extends beyond the immediate application. The Engine as a tool, empowers internal teams to support a sustainable  AI/ML capability, providing an additional competitive edge in a highly competitive and rapidly evolving logistics industry. 

Wrap Up 

Using the AI & Analytics Engine, the Client is able to build a repeatable collaborative ML development workflow for their data science team. This empowers data practitioners to iterate models and projects fast for continued experimentation and innovation efforts. Read on for more AI/ML industry solutions


Get in touch with us to find out how you can use the AI & Analytics Engine in your workflow for a predictive modelling solution! 

Contact Us

Similar posts

Subscribe and get the latest news, articles, and thought pieces direct to your inbox.