Excel scientific and engineering cookbook
Celebrity chef’ing by numbers
Book review Most people have a schizophrenic attitude to Excel, seeing it as “trivial” because they can use it with almost no training, but too difficult to use on more meaty problems. By way of response, David Bourg delivers a set of recipes that cook up some increasingly non trivial problems.
The cookbook style does however mean that rather than just telling you things like “to get log scales do this” the author does tend to assume that this is a problem for him to solve for you. But overall it works because almost no one will ever read this book from cover to cover, instead dipping into it achieve specific goals.
The section on Excel graphs is more than competent, and possibly is the basis for those who need help in visualising relatively dry equations without having to master more complex and expensive tools like MatLab and Mathematica. The introduction to finite difference methods is easily the clearest I have ever read, and possibly justifies the whole book if that topic is proving hard for you to master. I am also taken with the way that Bourg illustrates data manipulation with real data sets that you can download, and I can’t recall any other author dealing with the real problem of “weird” characters in imported data. This is used to take us through a range of real world statistical analysis beyond the precooked recipes in Excel.
What shines through each example is how Bourg brings in his experience of doing this in real life. If you’re doing serious number crunching in Excel, then this book can shortcut otherwise painful learning experiences. A small example is that most people seem to think you can’t drive Solver from VBA, and automating can make many tedious problems simply go away.
However the focus is on doing almost everything through no procedural sheets. That fits the way most Excel users think. Rather than VBA loops, he presents the unusual approach of ranges allowing you to look “inside” the calculation. Thus you can see things like Euler’s method go very wrong, rather more graphically than in Duffy’s book. (link to my review) He uses the Solver to deal with several issues that one would normally code iteratively, which makes it both faster and usually more robust. The code he does present is mostly quite good, except for his use of VBA integers. These little 16 bit wretches are a source of many irritating Excel bugs, and it’s not good that he promotes their use, when longs are available.
You can just about learn elementary VBA here, but you’d be better off with Walkenbach or Jackson & Staunton if you’re targeting investment banking work, since Bourg’s engineering background makes this a better entry-level engineering text than a finance one.
O’Reilly have done away with code disks, and replaced it with a free trial of their Safari online bookshelf. This is a pretty good system, including the full text of the book, but does mean that if you don’t get all the stuff you need from the book 45 days, you end up cut off.®