Re: [R] R and Google Code-In
Virgilio Gómez-Rubio
I am Cc'ing the GCi mailing list.
> Do you think that doing basic maps with Google and R might be a
> worthwhile activity, please?
Yes, that might be a good task and it could fit the following group of
activities:
5. Research: Tasks related to studying a problem and recommending
solutions
However, the tasks should be doable with R but the resulting maps could
be exported to Google Maps. If you have an interesting data set that
could be good. If not, I have location of tornados in the Contiguous
States (from NOAA, I believe) and we could ask the students to look for
the most likely location of Dorothy's place in Kansas (i.e., the point
with the highest intensity). It will require a bit of kernel smoothing
and mapping, which may be hard for a student but the mathematics behind
it are simple.
What do you think?
Best,
Virgilio
Erin Hodgess
I have the Houston monthly crime data. That could be fun.
Thanks,
Erin
2011/11/8 Virgilio Gómez-Rubio <virgili...@uclm.es>:
--
Erin Hodgess
Associate Professor
Department of Computer and Mathematical Sciences
University of Houston - Downtown
mailto: erinm....@gmail.com
Paul Hewson
visualising these kinds of data in school? You can get such data
visualised in places like New York Times,http://www.ons.gov.uk/ons/
guide-method/understanding-ons-statistics/interactive-content/
interactive-content.html). The attraction is that you instantly have
interaction with the data, with no software overhead and thus can do
this in any class where it's sensible to use data.
(arguing against myself, I always want to do this in R because I can
change how it works, what data it uses, exactly what we parameterise
and so on, and I'd love the mapping to be so easy that a high schooler
could do it).
On the other hand, R is great for (say) programming up a randomisation
test, which is a great way of understanding hypothesis tests. Might
not visuals and programming around threshold concepts like that be
really useful. Oh, and R is a great Bayesian statistics calculator,
so no-one from high school onwards need have an excuse for not using
Bayesian methods;-)
Best
Paul
Paul Hewson
Associate Professor
Plymouth University
On Nov 8, 4:33 pm, Erin Hodgess <erinm.hodg...@gmail.com> wrote:
> Sounds great!
>
> I have the Houston monthly crime data. That could be fun.
>
> Thanks,
> Erin
>
>
>
> > Dear Erin,
>
> > I am Cc'ing the GCi mailing list.
>
> >> Do you think that doing basic maps with Google and R might be a
> >> worthwhile activity, please?
>
> > Yes, that might be a good task and it could fit the following group of
> > activities:
>
> > 5. Research: Tasks related to studying a problem and recommending
> > solutions
>
> > However, the tasks should be doable with R but the resulting maps could
> > be exported to Google Maps. If you have an interesting data set that
> > could be good. If not, I have location of tornados in the Contiguous
> > States (from NOAA, I believe) and we could ask the students to look for
> > the most likely location of Dorothy's place in Kansas (i.e., the point
> > with the highest intensity). It will require a bit of kernel smoothing
> > and mapping, which may be hard for a student but the mathematics behind
> > it are simple.
>
> > What do you think?
>
> > Best,
>
> > Virgilio
>
> --
> Erin Hodgess
> Associate Professor
> Department of Computer and Mathematical Sciences
> University of Houston - Downtown
Dianne Cook
Unfortunately, R was not chosen as an organization for this year's GCI! I just learned this yesterday.
I was too slow getting projects organized. We will aim to start earlier next year, and be ready for the google announcement.
cheers,
Di
---------------------------
Di Cook
vis...@gmail.com