Building a data pipeline to predict damage to ship’s cargo

in Research, Sponsoring by
The goal of the student thesis was to conceptualize a data pipeline combining AIS data with environmental data measured by sensors to track the route of cargo and the conditions it has been exposed to. 

We support students and researchers by offering access to the FleetMon API Suite and our extensive AIS Data Archive with historical vessel position and port call data. Read this guest article we received by Niklas Scherer, a master’s degree student of the University of Applied Sciences in Bingen, Germany.

The academic project investigates a correlation between specific weather conditions a vessel was exposed to and occurring cargo damage. AIS data and weather data were used to examine if certain weather conditions on maritime high-traffic lanes are likely to cause damage to freight in order to prevent damage by realistic forecasting.

What happens to my cargo on its trip?

In a globalized world, products and materials are shipped worldwide on a daily basis. This poses significant challenges due to different climates, unpredictable weather conditions and ever so slightly changing routes which can easily mix up all of these variables. The goal of my thesis was to conceptualize a data pipeline that could combine AIS data with appropriate environmental data measured by sensors to track the route of cargo and the conditions that it has been exposed to. 

Proposed infrastructure and test run

To achieve this goal, AIS data was used to track a ship on its route from Germany to Australia. This long journey guaranteed to record a variety of temperature levels and many opportunities for sudden weather changes. The cargo container was fitted with multiple sensors to collect data on temperature, humidity, and acceleration in all directions. 

FleetMon’s API was polled in regular intervals to receive the ship location at a given timestamp, which could then be automatically fed into the sensor interface. An external weather API was also used to receive data on maritime weather conditions, e.g., wave height. All of these were combined to create data points containing environmental conditions for each recorded AIS position for the entire trip to Australia. 

Results and potential use cases

The test run produced promising data on how the different variables changed over time as a function of the current position. Examining a single trip as a proof-of-concept does not allow drawing any conclusions, however, the used methodology proved to be sufficient to scale up the operation and collect data on many different routes. 

Once enough data has been collected, a follow-up project would aggregate and analyze it. Possible use cases include data visualization or even In the future, artificial intelligence could be used to make decisions on the best route based on the route and environmental conditions it has been exposed to, to protect cargo.