Pig Data: analytics for the Swiss swine industry

So far, Big Data has had little impact on pig and pork production in Switzerland. This project explores ways of using Big Data methods for making pig farming in Switzerland more efficient and improving animal health and welfare by using data from across a complete pig production supply chain.

  • Portrait / project description (completed research project)

    Dropdown Icon

    Key partners in all areas of pig production (farmers, veterinary surgeons, marketing organizations, animal feed manufacturers and slaughter companies) collaborated in the project and made their data available for analysis. Data included health, production, reproduction and nutritional data, mortality statistics, veterinary records, transport data, carcass quality and weather data. The interdisciplinary research team created a data store, which amalgamated this heterogeneous data. Methods for transporting, loading cleaning, processing and analysing the data were designed and implemented. Analyses focused on answering research questions that were identified by the pig production partners as being important for solving production problems. Results of the analysis were communicated regularly to the pig production partners who evaluated the results and provided inputs to improve both the understanding of the data and the results of the analyses.

  • Background

    Dropdown Icon

    Large data volumes can no longer be analysed using conventional methods. They have to be made usable by applying new methods. These methods have not had any impact on animal husbandry even though they would be of great interest for pig farming in Switzerland in particular. Swiss pig production differs from the intensive production systems in other European countries because of its complex, small-scale structure. Although all stages of production generate animal health data, it is not being used in a way that brings together all the various stages. If this information was suitably prepared and analysed, it will be possible to recognize new links, causes and risk factors in relation to diseases and/or a drop in performance, and to identify the best strategies for combating them.

  • Aim

    Dropdown Icon

    The overarching goal of the project was to develop new methods aimed at gaining a better understanding of, and optimizing, the structure and complexity of the pig farming and production network in Switzerland. To achieve this goal the project aimed at:

    1. demonstrating that new, useful and valuable information can be created by combining many disparate data sets from across a complete pig supply chain
    2. developing trusting relationships between the researchers in the project and the data collectors in a complete Swiss pig production system so that they are comfortable enough to share their data and the results of data analyses with each other
    3. processing and joining the many different data sets from a pig supply chain together in a single data store that is suitable for data analyses
    4. using existing methods or developing new methods to analyse the data in order to produce new information that is useful and valued by the swine supply chain partners in the project
    5. successfully conducting a research project involving researchers from different disciplines who have never worked together before
  • Relevance/application

    Dropdown Icon

    The adoption of Big Data approaches by the Swiss swine industry has been relatively slow, especially when compared to other industries and countries. It was well known that the structure of the swine and pork production system in Switzerland differs greatly from the intensive production systems in other countries. For that reason, the methods developed and applied elsewhere were of limited value for Switzerland. It required to create a new strategy for the adoption of Big Data approaches for the Swiss swine industry. This project demonstrated that it is possible to combine data from various actors in the pig supply chain. It also firmly established that this approach can generate new and useful information. The Pig Data project can be seen as a pilot experience about the application of Big Data approaches in the Swiss swine industry. Its success was key to drawing the attention of key stakeholders (such as the Federal Food Safety and Veterinary Office) to the importance of this topic and for the planning of new projects that will further explore the field. This is evidenced by the official support of a new project that aims at establishing a ‘competence & information centre for pig health in Switzerland’ which will collect, analyse and publish health data from various projects and practitioners in the field in order to obtain a real-time overview of the population’s health status and the occurrence of emerging and re-emerging diseases.

  • Results

    Dropdown Icon
    1. The project demonstrated that new, useful and valuable information could be created by combining many disparate data sets from across a complete pig supply chain. The pig data supply-chain partners reported that the information created to answer the Dream Queries was interesting and useful. In addition, the supply-chain partners recognised that the information could not have been created without combining data from many data providers who normally would not have combined their data.
    2. The project successfully developed trusting relationships between the researchers in the project and the members of a Swiss pig production system. The pig supply-chain partners shared their data with the research team, they collaborated with the research team throughout the project, they allowed the results of the analyses of their data to be shared among all supply-chain partners and they approve the publication of these results in scientific journals and at scientific and industry conferences.
    3. The research team successfully developed a central data store and developed methods to clean, process and join the many different data sets from a pig supply chain. The cleaned and joined data were made available to the research team who used them for analysis to answer the dream queries.
    4. The research team used existing analytical methods and developed new methods to successfully analyse the data to create new information that was valued by the supply-chain partners in the project. New methods were developed by the graduate students in the research team and made publicly available through publication in scientific journals.
    5. The research team was very diverse which created a challenge for agreement on the best research strategies. Developing consensus on the best approaches was achieved through frequent online and in person meetings often difficult. The team was able to jell into working unit that achieved the goals of the project.
    6. The project was successful at developing new methods and creating new information that was accepted by the scientific community. The project team has published 3 studies in scientific journals (5 are currently being written), 2 studies in scientific conference proceedings, and presented at 9 scientific conferences.
  • Original title

    Dropdown Icon

    PIG DATA: Health Analytics for the Swiss Swine Industry