[R-SIG-Finance] R-Finance Tutorial

Gabor Grothendieck ggrothendieck at gmail.com
Fri Jul 2 05:50:00 CEST 2010

On Thu, Jul 1, 2010 at 10:23 AM, Jeff Ryan <jeff.a.ryan at gmail.com> wrote:
> With respect to what Brian said earlier, one of the powerful aspects
> of R is that there isn't 'one way' to do something.
> Depending on your needs, solutions come in different packages.  The
> downside to this is that it become a challenge to not learn the
> language per se, but more the 'best practices'
> To get the answer to that, or part of the answer at least, the best
> resource has to be the archives of R-SIG-Finance.  Additionally the
> papers/presentations from R/Rmetrics Meielisalp (going on this week in
> Switzerland for the 4th year...) and those from the 2009 and 2010
> conferences in Chicago www.RinFinance.com
> If you are looking to not reinvent the wheel, take a look there for
> inspiration.  Also I have quite a few slides/presentations that will
> give a good overview of things from an infrastructure perspective,
> albeit biased toward the packages I cover. (quantmod, xts, IBrokers,
> ...)
> http://www.quantmod.com
> http://www.quantmod.com/examples
> http://www.quantmod.com/Columbia2008/
> http://www.quantmod.com/Rmetrics2008/
> http://www.quantmod.com/Rmetrics2009/
> http://www.quantmod.com/Vienna2009/
> For time series, do read the vignettes on zoo and xts, as they will
> give you a taste of what both can do.  There is also a book about the
> time-series classes from Rmetrics that is free (in addition to some
> very nice ones that cost a little money).  The book isn't completely
> unbiased of course, but more information is always better.

More information is better but misinformation is not better.  I assume
that the material in the free book is accurate relative to rmetrics
since it was written by the rmetrics developers but it has basic
misunderstandings and mis-statements relating to the design and
functionality of zoo. If you want to learn zoo properly you need to
read the zoo vignettes and help pages.  For example, the book in
question states that zoo does not handle Financial Centers but that
confuses the time series class with the time class.  zoo, the time
series class, does not need to handle properties of the time class
since the time class handles those.  The zoo user can certainly use
Financial Centers by using the appropriate time class.  This requires
no development or customization.  It just works right out of the box.
A second option is that the zoo user can use zoo via the xts subclass
of zoo (in the xts package) which in turn can handle timeDate and its
Financial Centers.  A fairer comparison might have instead pointed out
that zoo handles just about any time class so that you can use the
appropriate class to the problem.  Some of the other comparisons are
misleading in other ways such as showing a black and white graph from
zoo vs a color graph from rmetrics instead of pointing out that zoo
supports significant portions of the vast array of both classic
graphics and lattice graphics and xts and quantmod have further
graphical facilities.  Also some of the comparisons are actually false
whereas certain others were true (zoo has been developed over  many
years) but not now.  Yet other statements are true but immaterial
since they relate to features that are basically never or almost never
used so they are not really a consideration.   Undoubtedly not all of
it is false and misleading but enough is that it can't be recommended
for learning zoo.

More information about the R-SIG-Finance mailing list