Proposal: Adding an FDE Section to the TRL

Florian_Vogt

Hi all,

 

I have been collecting ideas, questions and answers, some use cases and open questions also on FDE. The current thread on more background to FDE (https://community.jmp.com/t5/JMP-SE-Discussions/FDE-maths-behind-fitting-options/m-p/662515#M6331) also indicates interest in the topic.

After selling JMP Pro for FDE once, I encountered some difficulties throughout the year. One is definitely that FDE (FDA in general) appears to be widely unknown (I can mainly speak for the chemical customers). It is very easy to spark interest of prospects, but actually selling it remains a challenge (to me at least ). I think it would be very helpful to have a better understanding of what it can deliver, how it works etc. to promote it better and more often with customers.

My proposal is a location where we collect use cases, questions and answers, day-in-the-life-before-and-after-stories, and additional content like theoretical background or training materials (for SE education) @Di_Michelson  .

As a starter: One use case was 'semi automated peak fitting using Python'.

Situation before:	Scientists are collecting many samples over long periods of time. Measurements resulted in a spectra like curve, which looked very different in shape and number of peaks, depending on sample characteristics. The scientist were visually assessing the different shapes and tried to fit the curves shape characteristics manually. The issues were different numbers of peaks, which had to be identified first. Once analyzed, the output was not really easily understandable. A data scientist could analyze the data in Python, but this was not scalable, as it was at the cost of his own working progress.
Situation after:	Analyzing the Data in FDE showed instant results (really within 2 minutes), because the peaks and shapes were automatically analyzed and linked to the characteristics of the sample. Just by using the Profiler, they could see what they were after immediately. Without need for training, the scientist could immediately relate the insights to the characteristics of the respective product.
Personal Wins:	Better insights faster, better development outcomes Easier to interpret and communicate results No need to switch between software packages No need to wait for expert/data scientist
Business Wins:	Better insights faster No need for data scientist consulting More scalable solution Standardized output One software for many tasks

 

I am curious what you think of this idea and what additional thoughts you have!

 

Best regards from Heidelberg, Germany

Florian

malcolm_moore1

Thank-you for starting this discussion Florian. I think it would be great to develop a library of stories along with frequently asked questions and answers. As you say, not many of our users are familiar with FDE and they often need more assurance and details regarding the method than we are currently able to provide.