Stretching my brain back a decade, I used to work a lot with fourier transforms (FT) which are foundational methods of analysis MRI / NMR data. FT is so powerful and versatile that is widely used across all physical sciences to extract signals.
It is a technique that keeps flashing across my radar. A lot of very practical data analysis is about ways we might identify key components of the data, or isolate underlying data-generating processes and trends. I often wonder whether the FT might be useful. It isn’t something we see much of in the analytics world, or in statistics (with the except of time series analysis).
I am going to leave this post here for now. This is the start of an idea. Potentially a very large and interesting project if I can find the time for this. This post will be a little like a notebook, and if I do start to pick away at this I will add interesting links below to follow up on.
Follow up links:
“Joseph Fourier showed that any periodic wave can be represented by a sum of simple sine waves. This sum is called the Fourier Series.” FT tutorial with R