Data Analysis in the Earth Sciences Using MATLAB

Rating:
Summary: Short in Theory Description
Review: Being a graduate student, I don't much like this book. The theory description is not nough, and there're some mistakes in the Matlab programs. For the propose of partice and technique application, this book is not bad, but for those students really want to know much about data analysis, this book is not quite enough.

Rating:
Summary: matlab - not just for earthies
Review: The title of this excellent book is inappropriate. It should be "Data analysis using Matlab" or perhaps "Data Analysis in the Physical Sciences using Matlab", since the material is quite general and the book will be useful to many folks outside Earth Science.

Matlab is fast becoming the lingua franca for data analysis in many branches of physical science and engineering. How far it will spread beyond this domain is a question. It remains to be seen whether biologists, psychologists and the like, will give up their point/click interfaces. And at the other end of the scale, high-level users will continue to eschew matlab for more application-specific languages (e.g. S+ and R for statistics) and more efficient languages (e.g. fortran for ocean-modelling runs that take months of CPU time). However, there is a vast range of users who find that matlab is a tool that fits the hand well. You'd be hard-pressed to find a physical-science laboratory without well-thumbed Matlab manuals scattered among the fortran manuals and the tables of integrals.

Until recently, the problem with matlab has been a lack of good books telling how to use it in real-world applications. Matlab usage has spread mainly by word of mouth. That's terrific for high-level researchers since any form of communication (even about silly syntax errors) may very well lead to fruitful collaboration. But it's not so good for students, who need texts that start from the ground up, using matlab at every stage. Why should students learn data-analysis techniques from textbooks that are "language-neutral", when they are most likely to be using matlab when they enter the research community? Wouldn't it make more sense for them to learn about fast fourier transforms, fractals, etc., by actually doing computations instead of reading about computations, by actually drawing graphs instead of looking at them?

All of this makes me wonder why the market hasn't been flooded with great books about using matlab in science. But it hasn't.

Middleton's book thus fills an important gap. In it you'll read a bit about probability, regression, principal component analysis, cluster analysis, time series, spatial mapping, fractals, etc. In other words, Middleton touches on many of the tools and techniques that are used daily in the physical sciences. The key is that he gives matlab examples of all that he does. (As you'd expect, these are on an enclosed floppy disk.)

His approach is long on example and short on proof, and many readers will like this. I am a professor, and I'm not sure that I'd be happy using this book alone in a course, because of the "short on proof" aspect. But I may be an old fuddy-duddy. And, it must be admitted that a lot of science today is of the "hey, let's try this cool technique in our field" variety. (Who, of a certain age, has not seen catastrophe theory, chaos theory, fractal theory, wavelet theory, etc., wash over our fields like waves upon an unwet beach?)

I am certain that Middleton's book will find a comfortable place on my bookshelf, right near my elbow, and that I'll be recommending it to students and to fellow researchers. It's not complete, and it's not as deep as I'd like, but it covers a lot of ground and it does it in a highly practical way. But it will get you up and running quickly, letting you focus on applying data-analysis techniques to your work and not on developing the techniques. Hey, even the point-click crowd would do well to glance at these pages!

Rating:
Summary: matlab - not just for earthies
Review: The title of this excellent book is inappropriate. It should be "Data analysis using Matlab" or perhaps "Data Analysis in the Physical Sciences using Matlab", since the material is quite general and the book will be useful to many folks outside Earth Science.

Matlab is fast becoming the lingua franca for data analysis in many branches of physical science and engineering. How far it will spread beyond this domain is a question. It remains to be seen whether biologists, psychologists and the like, will give up their point/click interfaces. And at the other end of the scale, high-level users will continue to eschew matlab for more application-specific languages (e.g. S+ and R for statistics) and more efficient languages (e.g. fortran for ocean-modelling runs that take months of CPU time). However, there is a vast range of users who find that matlab is a tool that fits the hand well. You'd be hard-pressed to find a physical-science laboratory without well-thumbed Matlab manuals scattered among the fortran manuals and the tables of integrals.

Until recently, the problem with matlab has been a lack of good books telling how to use it in real-world applications. Matlab usage has spread mainly by word of mouth. That's terrific for high-level researchers since any form of communication (even about silly syntax errors) may very well lead to fruitful collaboration. But it's not so good for students, who need texts that start from the ground up, using matlab at every stage. Why should students learn data-analysis techniques from textbooks that are "language-neutral", when they are most likely to be using matlab when they enter the research community? Wouldn't it make more sense for them to learn about fast fourier transforms, fractals, etc., by actually doing computations instead of reading about computations, by actually drawing graphs instead of looking at them?

All of this makes me wonder why the market hasn't been flooded with great books about using matlab in science. But it hasn't.

Middleton's book thus fills an important gap. In it you'll read a bit about probability, regression, principal component analysis, cluster analysis, time series, spatial mapping, fractals, etc. In other words, Middleton touches on many of the tools and techniques that are used daily in the physical sciences. The key is that he gives matlab examples of all that he does. (As you'd expect, these are on an enclosed floppy disk.)

His approach is long on example and short on proof, and many readers will like this. I am a professor, and I'm not sure that I'd be happy using this book alone in a course, because of the "short on proof" aspect. But I may be an old fuddy-duddy. And, it must be admitted that a lot of science today is of the "hey, let's try this cool technique in our field" variety. (Who, of a certain age, has not seen catastrophe theory, chaos theory, fractal theory, wavelet theory, etc., wash over our fields like waves upon an unwet beach?)

I am certain that Middleton's book will find a comfortable place on my bookshelf, right near my elbow, and that I'll be recommending it to students and to fellow researchers. It's not complete, and it's not as deep as I'd like, but it covers a lot of ground and it does it in a highly practical way. But it will get you up and running quickly, letting you focus on applying data-analysis techniques to your work and not on developing the techniques. Hey, even the point-click crowd would do well to glance at these pages!