Within the amount of data that hatchery producers and fish farmers’ record are hidden patterns of causal relationships. Currently, producers aren’t using these continuous records for anything but daily operations. Analysis of production data is complicated and requires tools which are more sophisticated than what is currently available.
However, the computer program FindIT aims to use production data strategically, both in analysing own production and as a tool for benchmarking.
What is FindIt
FindIt is a unique tool for storing and analysing production data, and benchmarking. The system was developed primarily for the analysis of data from hatcheries, but can be connected to production data until harvest. Data from many producers are submitted to the system and each farmer can use the system to find new connections in the production performance. Data from other producers submitted to the database helps in this process.
A prototype of the system is now available and a commercially available, license-based system will be on the market in two years.
Grete Bæverfjord, a researcher at Nofima in Norway, will soon be inviting Norwegian fish farmers to join in.
“Today a lot of the knowledge are only in the heads of those who currently work on a site, while the data foundation is hidden in files and binders that nobody uses. As production is scaled up, it is impossible to keep track without good tools. We are looking for lasting knowledge that can be generated on the basis of the data that is already recorded on the sites,” says Bæverfjord.
“It is important to emphasize that the system is completely secure,” says Bæverfjord.
This means that any participant can see and work with their own data, but when a participant analyses the underlying connections, she will also be using information from the data we receive from others, but without knowing the identity of the data. This knowledge can be used to understand the past, manage the present and predict the future,” she explains.
A tool for data mining
The idea for the tool came through the work of the EU project FineFish, which specifically looked for causes of skeletal deformities in fish farming. There was little information to find on the causes of deformities from the producers’ own production data, which was an eye-opener also for the producers.
With this as a basis, the Belgian company Pepite was commissioned to develop a prototype analysis tool where variation in production data and production results taken over time is used to find underlying connections that are not apparent otherwise. The solution was ‘data mining’ and a tool called FindIt.
Data mining is based on large amounts of data and uses computer power to find patterns in the data which the human brain does not have the capacity for. Data Mining is a discipline within computer science that combines a variety of techniques such as pattern recognition and statistical modelling.
The system requires large amounts of data, and the value of FindIt will increase with the number of participants over time. Goals for the performance in different production phases will be set, e.g. survival at first feeding, growth in different periods and superior percentage at harvest. These performance goals will be adapted to the species and type of production. By following these over time and simultaneously having the link to production conditions, it is possible to extract knowledge about what leads to good results and what leads to poor results.
A brand new product with a longer time range
“This tool will not be launched as an alternative to other software tools on the market. FindIt uses the same data, but by using data mining it is built on very different analytical principles.”
“Hence, it should enable quicker progress and increased profitability,” says Bæverfjord.
Grete Bæverfjord and Synnøve Helland in Nofima have taken part in the development phase and their role now is to implement the tool in Norwegian aquaculture. They are in dialogue with Norwegian project participants to adapt the system to Norwegian conditions.
“Once we have the system up and running it will be open to participants from outside the project group. We see that Southern European hatchery producers are coming forward to become active participants in the work that is on-going and there is already great interest from salmon producers in other countries.”
“For FindIt to be useful for Norwegian producers, we are dependent on getting access to Norwegian data so that we can ensure that the system is also adapted to salmonids, not only marine water species,” says Bæverfjord.
– Siri Elise Dybdal