Knowit detects illegal fishing activities using multiple data sources in SAS® Viya®

Illegal fishing is a big problem in the Northern sea outside of Norway. A team from Knowit wanted to find a way to detect fishing activities in illegal areas along the Norwegian coastline. The result was a model that can help authorities detect and put an end to illegal fishing. The Knowit team won the Nordic Hack In SAS® Viya® competition with their contribution.

The government policy in Norway is based on the principle on sustainable harvest with strict fishery regulations. After assessing different challenges where they could use the open AIS data (Automatic Identification System) from the Norwegian Coastal Administration, the Knowit team set to find an efficient way to detect illegal fishing.

“One of our team members is from the Lofoten islands in the northern part of Norway where fishing is a big industry. He knew of the challenges imposed on the industry due to illegal fishing as it can lead to overfishing, species endangerment and marine habitat destruction. We found out that this was a meaningful challenge that we wanted to solve,” said Lars Kvinge, Data Scientist at Knowit Decision, one of the four team members from the Knowit team who participated in the competition.

Open data sources combined with visual analytics

As the AIS data does not contain information on whether a boat is a fishing boat, ferry, coast guard or other types of boats the team needed to extract the information from naval websites. With RPA (Robotic Process Automation) the team gathered all the information they needed about the boats online. All the information was then added to SAS® Viya® Visual Analytics to analyze the pattern of the boat to detect if it was a fishing pattern or non-fishing pattern. The team used a machine learning model to train data for the model.

“With SAS® Viya® it was easy to identify fishing patterns. We could quickly see the difference between boats that were fishing and not. We used the data we got from Viya® as training data for machine learning in Phyton and sent the information back to the platform to automatically identify if the boat was fishing or not,” says Lars Kvinge. “The computing power and the service in SAS® Viya® is great when you have a vast amount of data,” he adds. 

The model connects historical data in an area of interest to predict and highlight the vessels that has a behavioral pattern that could indicate illegal fishing. The team is now working on connecting a real-time AIS-data stream where the result is a model that can automatically warn the coast guard if it dissects any illegal activities.

Global recognition at SAS Global Forum

The jury was impressed by the way the Knowit team demonstrated a solution to a problem that is essential for a long-term and sustainable society, and they went all the way to the top in the competition. A part of the first price was a trip to the annual SAS Global Forum, this year in Denver.  

“It was a great trip. We were a bit overwhelmed when our case was presented with over 500 people from the industry in the audience, but it was truly amazing to get that recognition. There were many inspirational presentations and it was an eye-opener to hear all the possibilities offered on the SAS platform,” says Lars Kvinge.