Shaping up to erroneous data, for greater impact
Charles Sturt University in Wagga Wagga wanted to explore how our Shaipup infrastructure can help them deliver their high-quality data analytics capabilities directly to industry.
The team explored data discrepancies between different animal weighing techniques, and managing erroneous data for more rapid and impactful outcomes.
Overview
-
High quality data analytics algorithms were developed by CSU researchers.
-
They were hosted on Shaipup to both protect the intellectual property and make it available to industry.
-
The algorithm can be accessed by non-technical users via a spreadsheet or directly by software integrators via an application programming interface (API).
-
The team was able to host and test their algorithm in a very short time window.
Project partners
The project
The problem:
-
CSU’s goal was to develop a cattle growth rate calculator to help farmers better plan for when their cattle would be market ready.
-
The team quickly realised that any predictive growth calculator will only be as good as the data it is given – garbage in, equals garbage out.
-
Ensuring the cattle weight data was robust and statistically relevant was a far greater challenge than initially anticipated, and also presented as another opportunity.
-
Good quality data is required for genetic improvement programs, quality assurance programs and market access, supply chain management, etc.
-
Each time a researcher collects data they apply statistical tools to identify erroneous data. These checks ensure the data analyses are based on robust input data.
-
However, this system of automated checking for erroneous data has applications for all of industry but has been difficult to share with cattle producers who also wish to identify potential erroneous results.
-
Increasingly data collection is being automated and automated checks are needed to make sure data quality is maintained.
-
The team explored data discrepancies between different animal weighing techniques, looking at the differences between automated in-paddock weighing and the more industry standard yard-weighing.
-
This problem turned out to be the crux of the issue – and represented the greatest opportunity.
The solution:
-
The CSU team worked directly with cattle producers, utilising the TerraCipher Trakka infrastructure, enabling farmers to contribute their data to the project data analytics team.
-
The team developed a data clean-up algorithm, using sophisticated statistical modelling, which the team then hosted on our Shaipup platform.
-
Shaipup then made this algorithm available to industry. Farmers could access it directly from a spreadsheet, while it was also made available to software integrators via an Application Programming Interface (API).
-
We worked closely with the CSU team to deploy the algorithm.
-
Our platform allows the researcher to interact directly with industry through their algorithms hosted as Shaips. These are available for non-technical users via a spreadsheet, or for software integrators via an Application Programming Interface (API). Check out CSU's Shaip in the marketplace.
Charles Sturt University's Outlier Detection algorithm publicly available in the Shaip marketplace.
Charles Sturt University's Outlier Detection Shaip being used via a spreadsheet.
Key benefits:
-
The Shaipup platform provides a wrapper layer that allows research organisations and industry experts to focus on developing insights through data analyses and associated algorithms.​
-
The CSU team was able to get the first iteration of the algorithm deployed and ready to be used in a matter of weeks (rather than months, or years).
-
Shaipup provides the ability for CSU to make their high-quality data analytics capabilities directly available to industry.
-
Shaipup allows CSU to rapidly share knowledge algorithms to a global marketplace, without the need for a complex licensing agreement with the IP behind the algorithm fully protected.
-
CSU can use algorithms, such as the erroneous data checker, to immediately generate revenue as well as providing meaningful usage statistics that can be shared with governments and funders.
-
The potential customers for the algorithm can easily access it and pass their own data in.
-
The CSU data checker algorithm is available for all producers to be able to pass their data in and get it checked before they start to make decisions.
Key benefits:
-
Checking livestock data can help lift the quality being used across the red meat industry.
​
-
The data analysis tools that researchers use to check data are high quality and can also be used to lift the quality of industry data.
​
-
Improved quality data that is standardised ensures improved decisions. It also ensures stakeholders that rely on the data or associated analyses can have confidence, for example using the data for genetic improvement or market access.