Perhaps a slightly different sort of science book, but I highly recommend "Foundations of Applied Mathematics" by by Humpherys, Jarvis, & Evans, the first volum

Perhaps a slightly different sort of science book, but I highly recommend "Foundations of Applied Mathematics" by by Humpherys, Jarvis, & Evans, the first volume of which was published in 2017, and of which the second supposedly will be published in 2019. Volume 1 is good overview of the Analysis needed in various applications, they give as examples "analysis, optimization, probability, statistics, differential equations, machine learning, and control theory". I found it useful as an overview of the theory one might need in a unified exposition, with a modern and pragmatic choice of topics and notation. Learning the topics previously from many books, all with different notations and ideas of pedagogy, can be confusing (to me at least), especially in understanding how different parts of the theory are interrelated. It even comes with computational "labs", the code for which is available freely online, something often lacking in books on numerical analysis, applied math, etc. My only complaint is that it somewhat leaves out differential equations. Definitely worth looking at! In particular the chapter on matrix calculus.