A cell is simply a computational variable (as opposed to the notion of variables in lambda calculus). A named range is a data structure (a struct). The rest are term rewriting
The defining feature of a reactive program is that relationships between inputs and outputs are automatically tracked, and changing an input will automatically update the dependent outputs.
Facebook had a whole experimental language (now abandoned) for writing reactive programs with imperative code: http://skiplang.com/
Having said that, other than Excel and some cute UI things, I've not seen much utility in reactive paradigms. Hence I am bearish on reactive programming. Someone please correct me.
And that's why when you start to use use VBA or js in Excel, it's a fail.
Excel power users (I am not one of them) know how to solve almost every problem they are presented to, using only Excel. Like SQL with relational data. With orders of magnitude better performance.
Main issue yet is scalability, when the dataset gets too big.