||New Reviews| |Software Methodologies| |Popular Science| |AI/Machine Learning| |Programming| |Java| |Linux/Open Source| |XML| |Software Tools| |Other| |Web| |Tutorials| |All By Date| |All By Title| |Resources| |About||
Keywords: Excel, statistical analysis, data, spreadsheets
Title: Analyzing Business Data With Excel
Author: Gerald Knight
Verdict: Excellent. Recommended for Excel super-users.
Like it or not, and there are plenty who don't, Microsoft's Excel is probably the most popular data analysis tool in the world. Undoubtedly the best of the applications that make up Microsoft's Office package, it's feature-packed, easy to use and easy to extend using VBA and the built-in functions and tools. However, ease of use can have a major downside in that users blind themselves (and each other) with science. The world is awash with good-looking Excel workbooks that are riddled with errors, coded badly and using features without understanding what's going on under the hood. It's this problem that author Gerald Knight tackles in 'Analyzing Business Data with Excel', published by O'Reilly.
Where most books on enhancing or extending Excel take a feature-led approach, Knight turns things around and starts with the problems to be solved. And frankly it's this problem-oriented approach that really sets the book apart from most of its peers. Instead of saying 'here's this really cool bit of code that does x', this book says 'I've got some forecasting to do, how do I do it?'. And if the solution needs features x, y and z, then they get introduced along the way.
Obviously there are some features which are used throughout, such as the use of array formulas (a skill that in practice seems to elude most Excel users), statistical functions and the use of pivot tables and pivot charts. These get addressed up front, in the first couple of chapters, and then they are used throughout the rest of the text.
The range of business problems that are explored includes workload forecasting, modelling (including regression), queuing, optimisation problems, statistical process control and process monitoring. Each of these topics is addressed with a complete application, using a combination of cell formulas, charts and custom Visual Basic code. The applications are well structured, with clearly differentiated sheets for display, data and calculation. As a general way of working these sample examples are well thought out and really do show the reader how to make the most of Excel.
Additionally, the book also looks at some other more generic issues, such as problems with importing data, data format issues and some clear and useful advice on formatting and effectively displaying data.
There are a few niggles worth pointing out. The first is that there is frequent use of statistical procedures in the book, and often these are shown as Excel formulae. For those with a bit more statistical knowledge it would have been nice to see these formulae in more traditional math notation as well. Secondly there are known issues with some of Excel's statistical formulae, and a mention of these would have been a useful aside and an antidote to those who believe everything that a spreadsheet tells them. However, these are really quite minor points and don't much detract from the book overall.
For the data analyst who wants to really get the most out of Excel then there are few books that compare in depth and quality to this one. This is one title that we can recommend very highly indeed.