Functions are a bit more complicated. Michaelson states that a λ function serves as an abstraction over a λ expression, which isn’t that informative unless we take some time to understand what abstraction actually means.
Programmers use abstraction all the time by generalizing from a specific instance of a problem to a parameterized version of it. Abstraction uses names to refer to concrete objects or values (you can call them parameters if you like), as a means to create generalizations of specific problems. You can then take this abstraction (you can call it a function if you like), and replace the names with concrete objects or values to create a particular concrete instance of the problem. Readers familiar with refactoring can view abstraction as an “Extract Method” refactoring that turns a fragment of code into a method with parameters that explain the purpose of the method.
I think having a good understanding of Lambda calculus is a huge advantage for a data platform professional, as it gives you an inroad to learning data-centric functional programming languages (e.g., Scala, R, and F#) and neatly sidesteps the impedance mismatch problem with object-oriented languages.