[R] OT: Where's the new Tukey?

[Liviu Andronic]

On Wed, Jul 18, 2012 at 12:30 AM, Kjetil Halvorsen
<kjetilbrinchmannhalvorsen at gmail.com> wrote:
> Venables & Ripley:
> "Modern Applied Statistics with S (fourth Edition)"
>
[..]
>
> On Sat, Jul 14, 2012 at 4:01 PM, Larry White <ljw1001 at gmail.com> wrote:
>> I'm looking for a single book that provides a deep, yet readable
>> introduction to applied data analysis for general readers.
>
In my experience MASS doesn't apply to general readers. More to experts.


> I'm looking for coverage on things like understanding randomness, "natural
>> experiments", confounding, causality and correlation, data cleaning and
>> transforms, lagging, residuals, exploratory graphics, curve fitting,
>> descriptive stats.... Preferably with examples/case studies that illustrate
>> the art and craft of data analysis. No proofs or heavy math.
>
I'm no expert, but I'm very happy with what I'm reading in
'Statistics' by Freedman et al. (2007) [1]. This book concerns itself
with providing the reader with a clear, intuitive and accessible
understanding of the fundamentals of statistics. Its hallmark (for
better or worse) is the thorough avoidance of formulas or
incomprehensible math jargon. (It still contains a lot of proper
jargon, but it doesn't assume, as many books do, that the user
perfectly understands all the mathematical and statistical terms.) The
book requires, essentially, no prerequisites from the reader. As far
as I go, very good.
[1] https://en.wikipedia.org/wiki/David_A._Freedman_(statistician)

For a more rigorous, mathematic and advanced approach I like Applied
Regression Analysis and GLM by Fox (2008). Whereas the first book is
concerned with intuition, this book is focused on application (in the
context of regression analysis). To freely quote the author, the text
is as accessible as possible without the material being watered down
unduly. The prerequisites for reading the book are higher. As far as
I'm concerned, the material is clearly exposed and the author tackles
head-on a lot of thorny issues that other books leave untouched. I
guess this qualifies as "deep, yet readable introduction".

This last book can be perfectly complemented with An R Companion to
Applier Regression by Fox and Weisberg (2011). This is to teach people
R in the context of regression analysis.

Regards
Liviu