CloudCal launch & XRF.guru redesign

[Lee Drake]

Hi all,

First, thanks to those who have been beta testing CloudCal for the past few months. The software has been much improved as a result. It now can import XRF data in native format from Tracers, Artax, and Elios. It also now implements machine learning - a weak AI can now identify optimal normalization, intercept, and slope coefficients for Lucas-Tooth & Forest calibrations. In the not too distant future, the software will be able to do calibrations from start to finish by itself. Though, the default design of the app is to make sure a human can review all changes. It has been humbling to automate part of my old job for Bruker. 

Anyway, CloudCal is now out of beta and is available for everyone to use. Note: It can read in PDZ files from classic Tracers with no problem. New Tracers have a slightly different format, which can occasionally cause problems when reading in data. If this happens to you, email me at b.lee...@gmail.com.

Alongside this, I am (finally) updating www.xrf.guru with a new design, integrated site search, and supporting info for the Tracer 5i. It will also be a public access point for apps, including:

• CloudCal - software for creating empirical calibrations
  
• Analyzer - basic PCA & spectrum matching app
  
• 5i Results - simple app for parsing the stored results from a Tracer 5i
  
• Cezanne - mapping & PCA app for Tracer & Elios
  

So, you can run CloudCal by following the instructions at https://github.com/leedrake5/CloudCal, or by using the online version at www.xrf.guru. I also put up a demo version with pre-loaded data, so that if you just want to practice making corrections, you can do that too. 

For those who use CloudCal in their research, you can cite the GitHub page & use its doi:

Drake, B.L. 2018. CloudCal. GitHub. https://github.com/leedrake5/CloudCal. doi:10.5281/zenodo.1251965

I will be writing up its mechanics in a publication, but it will be a year or so until it is out. The goal of the apps isn't just to make calibrations easier and more open, but also to push the technology forward. This is why it will be a playing ground for experimental calibration types, such as forest regression that is currently applied. To see the potential of forest calibrations, go to the demo page and click "train" for sodium (Na): http://www.xrf.guru/Apps/CloudCalDemo

If anyone has any questions, please let me know. In the meantime, I hope CloudCal is helpful. 

Thanks,

---

B. Lee Drake

Department of Anthropology
University of New Mexico
(505) 510.1518
b.lee...@gmail.com