About Me

My name is Christopher Mertin. I attended Florida State University from August 2010 through May 2015 where I earned a Bachelors of Physics and a Bachelors of Applied Mathematics. While attending FSU, I was able to do research in the John D. Fox Accelerator where I was able to not only assist with Nuclear Physics experiments, but also conduct my own. These experiments produce a lot of data, leading to a lot of data analysis projects to understand the process inside stars.

The previous project, and my Bachelors Thesis, led me to pursue a Masters in Data Management and Analysis at the University of Utah where I specialized in Machine Learning algorithms and visualization techniques. While enrolled at the University of Utah, I also became involved in research where I implemented a new form of neural networks which are much more efficient at learning. This dealt with Hierarchical Matrices as being the backend of these matrix multiplications, reducing the computation time from \(\mathcal{O}\left(n^2\right)\) to \(\mathcal{O}(n)\). The learning rate also greatly improved, and the official write up can be seen in my blog post on the matter. I graduated in May 2017 with this degree, where as of today I am using what I’ve learned to provide insights into multiple industries.

From here, I am currently interested in using Data Science and Machine Learning to help reformat the healthcare industry by making healthcare not only more affordable, but also more efficient, accurate, and cheaper for patients. This can also be used to reduce patient time in recovery for surgeries and care by letting doctors and insurance companies know the best procedure for a given case.

This website, datastat.io, is a host of my personal projects from data that I’ve taken off of the web and made projects of, utilizing statistical tests and machine learning with data analysis. There are sites that provide basic patient healthcare data that I want to work with, but the real value comes from non-anonymized data. However, the techniques utilized in this blog can be used for such cases.