Be Sociable, Share!

• 

Data Mining for Business Intelligence: Concepts, Techniques, and Applications in Microsoft Office Excel® with XLMiner® Book Review

I recently met Dr. Galit Shmueli online, and she informed me about this new book (now second edition).  The book is priced and aimed as a graduate-level textbook for business majors.  Wiley was kind to send me a desk copy since I teach with the University of Phoenix.  Having read this book, I believe the explanations and content provide a good foundation in data mining, and though the chapters are organized like many such books, the approach is what I would expect for business intelligence applications.

I believe this book is outstanding, well-written and full of intuitive advice based on strong mathematics.  I like many of the one-liners I highlighted throughout the book on how to think about data mining.

In the past, I had been able to recommend a number of statistically-based books, which examine the algorithms.  I recommend this book (which has so many contributors too beyond the coauthors) as a valuable resource to people I know who want to provide more statistical depth in their application of SQL Server Data Mining. I believe this book could be used by undergraduates.  The consistent business focus provides many ideas on how to think about data mining value.

I enjoyed not just seeing familiar equations, but also seeing the common words and phrases from statistics being used to describe the data mining process.  And I’m happy to have the early chapters talk about what to do for data mining, and not just splash a CRISP-DM model and assume people would know what to do.  In my data mining presentations, I even take another step back and ask people in audiences about the scientific method.  I liked the data visualization chapter since I believe the software interfaces should and will grow in that area in coming years.

With the extension for this second edition into time series, this book covers most the algorithms from SQL Server Data Mining.  I am aware that one particular Amazon reviewer (“Lew”) was seriously disappointed both at the cost of the book, and more specifically that it did not address his predetermined favorite program R.  In responding to Lew, I advised that people who are professionals in this area need to consider all software available, both open source and commercial, and not have a predetermined “free is best” philosophy.  I have expertise both with SAS and Microsoft Business Intelligence, and as a consulting professional, I choose an obligation to my clients to be aware of various technologies.  XLMiner is the choice for this book’s authors, and though the book examples came from this technology, I believe the introductions and statistical explanations were vendor-neutral.

I also responded to Amazon reviewer “Jessica Jean” about the need for professional guidance.  I am sad for her that in her online class she does not have sufficient professional leadership to guide her into better mathematics.  People who read only one vendor’s documentation (whether it be from Microsoft or SAS or R fans) will only get only one set of perspectives (logically tailored to that technology’s implementation).  Serious professionals seek good guidance across technologies and also seek people like myself who provide both formal training and onsite guidance.  There are many books beyond just this one resource, and there are communities and conferences for people interested in building one another.  Data mining is an area of active research and active practice development, and no one book or article can encompass the newest and best applications (either from the technology perspective or from the machine learning algorithm direction).   Given the emerging nature of data mining, I have therefore provided many links on my data mining website marktab.net.

This textbook is a good introduction to data mining, and provides comprehensive information without being a several-volume tome or being 800 pages.  Especially business analysts would find the book practical.  Computer science majors sometimes take the topic at a more aggressive pace, delving either deep on the statistics or the algorithms (implementation) or both.  I believe those directions are important for people expecting “machine learning”.   By contrast, this book focuses on having practical datasets and applying them for specific problems.  People can access the datasets from the authors’ website and use not only XLMiner but any other data mining technology which may already be licensed and onsite.

If I had a critique (really a request for more) it would have been great to see this book in color (like Hastie).  Dr. Shmueli agrees with me and she said the authors worked hard on the graphical design of the content: I believe their effort shows.  She also mentioned that the instructor slides are in color, and I applaud their choice to include color.  In practice, we will be seeing more people experimenting with interactive visual analytics in color.

Be Sociable, Share!

•