NTNU alumnus Stian Ismar shares his experience working as intern at Camo Analytics this summer.
I am working as a software developer intern at Camo Analytics in Oslo for five weeks this summer. I recently graduated from NTNU (Master’s in Engineering and ICT) and was eager to learn more about how data analytics is used in the industry. I was especially curious about which methods Camo apply in their software.
The first three weeks at Camo Analytics has given me insight into how commercial software for multivariate data analysis is developed, and the broad range of fields and industries it is used in, such as pharma and agriculture.
The first three weeks at Camo Analytics has given me insight into how commercial software for multivariate data analysis is developed, and the broad range of fields and industries it is used in, such as pharma and agriculture.
The team at the Oslo office made me feel welcome from the beginning of the internship. Conversations, stand-up meetings and demos have exposed me to the variety of work that takes place at Camo, whether it be extending a feature in the software or developing client relationships.
My work consists of creating Python scripts to be used with Unscrambler, one of the software products Camo Analytics has developed for multivariate analysis. The latest version of Unscrambler enables the execution of Python scripts, extending the functionality of Unscrambler with everything Python has to offer. This extension is meant for two different user groups; domain experts relying on ease of use and data scientists with Python experience wanting to supplement their analysis.
Camo Analytics has developed a package called PyCamo which allows users to load datasets directly from Unscrambler 11 into a Python script through a dialog window. The package is also used to export the processed data back to Unscrambler from Python. Additionally, graphical user interface packages in Python enable user interaction with the scripts through simple dialog windows with drop-downs, radio buttons and text input. This means that the users can run the scripts with different datasets and input arguments efficiently, without necessarily having to see any Python code.
Over the summer, I am implementing various algorithms and methods in Python and connecting them to Unscrambler through PyCamo. The tasks are provided by Frank Westad and Torgrim Rønning at Camo. The projects so far have varied from programming blind source separation algorithms using independent components analysis (ICA) to scripting common performance metrics for classifiers using ROC-curves and confusion matrices. Other projects I have worked on relate to batch modeling, spectral baseline correction and PLS-DA.
It is motivating to work on the scripts knowing that they are valuable to Camo Analytics’s users. I am looking forward to the last week of developing Python scripts for Camo.
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.