The aquaculture industry in Norway is growing rapidly facing the challenge of managing vast amounts of data daily to ensure the optimal growth and health of fish stocks. Traditional data management approaches often fall short in handling the complexity and volume of this data, making it challenging for fish farmers to infer insights into their production operations. This work demonstrates the use of graph databases to visualize and analyze daily activities logged by fish farmers in the AKVA Fishtalk software [1], focusing on operations such as transfers, splits, and merges with fish groups. By employing custom algorithms to map these operations from data stored in traditional database management software onto a graph database, this work enables a holistic visualization of operations exerted on fish groups allowing a more comprehensive analysis that aids fish farmers in understanding the effects of various operations on fish product quality.
Our methodology centers on visualizing daily records stored by fish farmers in traditional database management software, e.g. the AKVA Fishtalk [1] in a graph database framework so-called Neo4j [2]. This involves the transformation of conventional data entries on stock, growth, feeding, and health metrics into a network of interconnected nodes and edges, representing the complex relationships and operations with fish farming activities. Custom algorithms were developed to identify key operational events, such as fish group splits and merges, and to visualize these within the graph database. Fig. 1 illustrates the Fishtalk software interface. In this environment, data is typically stored in a table format which is hard to understand and analyze.
Fig. 2 depicts a visualization from the Neo4j interface. This visualization is a result of deploying the custom algorithm that transfers the information stored in Fishtalk into a graph representation. This approach allows for a dynamic representation of the lifecycle and operational interventions for each fish group.
This work showcases the potential of graph databases in visualizing operations exerted on fish groups. Graph-based frameworks can provide fish farmers insights on how to optimize the various factors that influence the fish production process. Future directions include the development of predictive models to forecast outcomes of different operational strategies, thereby further enhancing decision-making processes.
References
[1] AKVA Fishtalk., Fishtalk. Accessed: Apr. 04, 2024. [Online].
Available: https://www.akvagroup.no/akva-fishtalk/
e: post@akvagroup.com
[2] Neo4j, Inc., Neo4j. Accessed: Apr. 04, 2024. [Online].
Available: https://neo4j.com/product/neo4j-graph-databas