By Geir Rune Flåten, Chief Solutions Officer
With the release of Unscrambler 11 we introduced support for Python scripting, giving users the best of two worlds. The Python extension to Unscrambler gives customers access to tap into the Python community. It’s a window to expand and explore methods, scripting opportunities and new algorithms, while at the same time benefit from a fully compliant framework and real-time platform to ensure quality, higher yields, shorter time to value and more profitable processes.
For multivariate analysis the methods available in Unscrambler have been refined during the last 35 years, but specific data or applications might require methods currently not supported in Unscrambler and the scripting creates a strong supplement to the Camo analytical platform.
Adding analytical value
What this means is that Unscrambler users can choose between thousands of free Python scripts covering everything from data import and pre-processing to modelling methods including machine learning, and execute them in Unscrambler with just a few clicks. The flexibility of running Python scripts together with the exploratory validation approach in Unscrambler, empowers users to develop and test models and algorithms for a wide array of analytical problem solving. Unscrambler users can also program their own or modify existing scripts if they want.
This is an extension and addition of analytical power to users of Unscrambler, and the Python support also offers organisations a platform to enhance and strengthen the internal collaboration between domain experts and data scientists.
Another aspect is speed for new analytical methods or innovation. Time-to-market is much improved for innovations by using the new scripting option and you can significantly speed up deployment of new software and new models.
Python scripting gives users the best of two worlds: Tap into the Python community, while at the same time benefit from a fully compliant framework and real-time platform to ensure quality, higher yields, shorter time to value and more profitable processes.
Bridging the gap between domain experts and data scientist
Python is a supplement to Unscrambler, and it serves two different user groups. Many users of Unscrambler are data owners working as domain experts close to a specific process in manufacturing, development or quality assurance. They rely on the ease of use and the maturity of Unscrambler. This user group is empowered to explore the feasibility of methods found in the Python or Camo communities.
Many large enterprises also have groups with data scientists that are analytical experts with Python as their native language. This extension to Unscrambler bridges the gap between these two different user groups, and this is important in order to speed up data-driven transformation in general.
Fully compliant and ready for production
Unscrambler is fully 21 CFR Part 11 and EU Annex 11 compliant giving solutions and models a proven and fast track to achieve compliance. If a Python scripted solution or model is being integrated with Unscrambler, the framework is inherently in compliance and you only need to qualify your script and its performance. Today you will always face an overhead of time and cost to secure compliance of a solution based on a Python script as you also need to consider the framework for the script.
Another big advantage of bringing Python into the Unscrambler framework, is the ability to run Python models on our Process Pulse real-time platform. Linking R&D and production on the same platform, accelerates the move from lab to pilot to production by offering seamless integration and scalability – with all the flexibility of Python included.
Crowd innovation
Organisations using Unscrambler, can easily get more analytical power with Python scripts. And we support it by arranging Python hackathons and encourage users to share and collaborate on the Camo community in an agile and responsive manner. We expect machine learning-based solutions from the Python Community to be integrated in Unscrambler, and we expect to see a growing collaboration between the everyday users of Unscrambler and the data scientists. This will hopefully constitute a happy marriage between the data owning engineers in the specific domains with the analytical powers of the data scientists.
Python ressources on the Camo Community >>
Read about the Unscrambler analytics suite >>
Talk to an expert
Get in touch if you have questions about our products, platform, how to get started or how best to address your analytics needs.