Full Stack Analytics
Performance improvements in JS VM’s along with the addition of async/await to NodeJS makes it a powerful platform for server-side development. I also find the ability to write-once deploy-anywhere in Typescript to be an additional selling feature. To illustrate the point, I’ve begun development of a Typescript library designed to move many of the frequently-used computations in R into Typescript. While this is largely a proprietary development, I have open-sourced a small portion of the code. It has been used to execute analysis detailed in the book ‘Machine Learning in R’ by Lantz on the client side in Angular 2.
I recently added another repo illustrating how the same type of analysis could be performed server-side in Node/Express. Only summary information is sent to the client. A detailed writeup of the process is available on the Github repo and I know you want to check out the source, so click on that link and enjoy!
This development illustrates the benefit of authoring in one language (Typescript) and being able to deploy either client- or server-side, depending on the specific needs of each application.