Project Department: Uni Research Computing (group: Center for Big Data Analysis) period: 01.05.16 - 31.12.17

About the project

Environmental and biological requirements & surveillance database for closed-containment systems

The CtrlAQUA DATABASE is a project part of the Centre for Research-based Innovation (SFI) CtrlAQUA consisting in the building of a complete and optimally designed database for managing, processing and analyzing a huge volume of research/production data from salmon farming in closed-containment systems (CCS), ingesting it into the Big Data computational infrastructure of CBDA (Hadoop, Spark, Hive, OpenTSDB, much more..) and resorting to cutting-edge Artificial Inteligence techniques. Examples of the kind of data being acquired are parameters and time-series about fish welfare and water quality in CCSs. The major aim is to make it available easily for further analyses, such as evaluating the environmental and biological requirements for post-smolt salmons and providing a surveillance database for closed-containment systems.

This means in practice to update and expand the current standards in salmon farming with new knowledge obtained within CtrlAQUA, fully exploiting the controllability provided by close-containment. The database will acquire as much data as possible from CtrlAQUA experiments and production trials from User Partners in order to maximize the usefulness of Big Data and Machine Learning techniques. The data volume that can be stored is huge, and the larger the better to determine production requirements, solutions, predictors, and general knowledge otherwise unattainable.

The power of Machine Learning enables to extract data from the database asking specific questions but exploiting it as a whole,  which allows to unravel new data representations and relevant parameters, tune the relative importance of parameters, learn features unknown from prior knowledge, and ultimately get new results in assessing environmental and biological requirements. Data from all inputs will be standardized using a master list of parameters and anthologies, make it possible to produce cross-partner analysis (eventually in a blind way), with much greater power to the interpretation and understanding.

In the long-term, the database can provide a platform to be used beyond CtrlAQUA's lifespan by new tools that could be implemented in software such as Aquafarm or FishTalk. These innovations would allow individual companies to use site specific production data linked to historical data, climate data, and the rest of the database data, predicting at best when the fish is optimally transferred from one situation to another, maximizing robustness, development, production and quality.

For more details about the project in its full context, see the official CtrlAQUA Annual Report for 2015.  

[a sketch of DATABASE, done with stock images from http://logomakr.com available through CC BY 3.0 Licensing]

cp: 2019-12-04 11:15:31