Advanced Excel for Scientific Data Analysis (英語) ペーパーバック – 2004/1/15
Kindle 端末は必要ありません。無料 Kindle アプリのいずれかをダウンロードすると、スマートフォン、タブレットPCで Kindle 本をお読みいただけます。
Excel is the leading spreadsheet both in its widespread distribution and in its computational features. In scientific research, spreadsheets are often used to organize and plot experimental results for reports and papers. Spreadsheets are also much-used tools in teaching some of the more quantitative aspects of science. This guidebook is different from the majority of existing Excel books in that it emphasizes the design of solutions to unique problems rather than simply the mechanics of spreadsheet use. Its focus is on the use of Excel to analyze numerical experimental data usually encountered in the physical sciences. The core of the book discusses the two primary approaches to scientific data analysis, least squares and Fourier transformation. Other cases in which experiments must be compared with the results of numerical simulations are also briefly discussed. Macros are presented as examples that readers can modify for their own purposes. The text is illustrated throughout with practical examples.
The book is nicely organized and includes detailed examples throughout (SUSANNE MAY, Division of Biostatistics and Bioinformatics University of California San Diego La Jolla, California, USA - Biometrics)
'...the book does what the author promises: it takes the reader "beyond the standard fare of Excel" and into a world of sophiscated applications that many scientific users of Excel should explore.' Journal of Chemical Education, Vol. 83, No.1 January 2006.
A couple notes may be in order.
1. He clearly prefers Excel 2003 and is quite emphatic that upgrading to 2007 is unnecessary and possibly undesirable, unless you have to. (2007 represents a syntax break from earlier versions, and the menus are even worse.)
2. He asserts that with his Macros and VBA Excel is a powerful general purpose language for data processing. He is also clear however that for REAL number crunching (Quantum Mechanics, Cosmology etc.) it will be way too slow, and you should use something heavier duty or a compiled language. For myself I intend to do everything I can on Excel (with his help) in the future, and upload data to Mathematica only when absolutely necessary.
There is very good explanation about fitting techniques, and a rough introduction, may be so light, to the Fourier transform and deconvolution. May be there is a need to more in deep explanation about non parametric least square fit (with splines for example), although in the book R de Levie explain Savitzky-Golay filter very well.
I not in agree to the comments of A. Scientist, you can use excel in science because vba has an advantage that matlab, R... doesn't at least at my knowledge: with vba you can call precompiled subprograms coded as DLL, originally code in C++ or Fortran (Financial Applications using Excel Add-in Development in C/C++). Using this feature any program is much quicker than any script in matlab o R. There are thousands of free codes in C++ o FORTRAN that the user only has to compile as DLL.
The only trouble is that this book do not mention this techniques.
I'm a physicist, so I can't comment on the chemical models presented in the book. Still, I learned a lot of things about Excel that I didn't know and it helped me implement customized VBA functions in some of my spreadsheets. In that respect, Chapter 8: Write your own macros, proved very useful. I got something else from the book: use Excel at your own perils. It is great for quickly whipping together a plot, a quick calculation, but it is not a robust numerical calcuation engine. The book advises to declare all variables as double precision - it is one way to alleviate the problem.