DEVELOPMENT OF THE DATA WAREHOUSE ARCHITECTURE FOR PROCESSING AND ANALYSIS OF THE RAW PIG PRODUCTION DATA

Abstract

Precision Livestock Farming (PLF) approach requires immense amount of data
collection, aggregation and processing, using various hardware and software
solutions, and is applied in many farms trying to achieve the most efficient and
effective way of production. Hardware and software compatible systems capable of
achieving this goal are called Farm Management Information Systems (FMIS), and
are a necessity for a complete and successful implementation for Precision
Agriculture (PA) branch approaches. However, most of commercially available
FMIS do not only focus on crop management, but also have limited availability to
small and average sized farms, in terms of price, supported language and specific
features. Simpler FMIS, on the other hand, do not have necessary capabilities to
fully support PLF. There are currently very small amount of high grade pig farm
oriented FMIS, especially for farms with less than a hundred of sows. Therefore,
there exists the need for solutions for managing farms with limited number of
sows. To help address this need, authors proposed and developed architecture for
unified data warehouse (DW), which was scalable and extendable cloud based data
storage and processing system with support of individual data analysis. DW has
capabilities to data interexchange and/or be integrated in existing FMIS throughout
variety of data-in/data-out interfaces, like UIs, unmanned data supplier or
consumer systems. The core of the DW is designed to provide data processing
flexibility and versatility, whereas data flow within the core is organized between
data vaults in a controllable and reliable way.