[R-sig-Geo] Disease Mapping & Spatial Statistics
chris english
englishchristophera at gmail.com
Thu Jun 11 11:40:28 CEST 2015
Hi Josh,
I find this forum very helpful as well and like yourself am new to and
learning R and spatial statistics. With that in mind I will leave some of
your specifics aside to those better qualified (for example your earlier
discussion with Dominik regarding box plots in ggplot) and describe a
process I use to better ground myself in both understanding and finally
executing what I want (hope) to accomplish.
Like yourself I generally have a specific that I am chasing, but over time
I've found it is less productive for me to chase it directly, and that I am
(eventually) better grounded if I start from a ten thousand foot view. I
still want to achieve my specifics, but I also want to understand
comprehensively what happened in the course of the process especially if I
am going to report/defend results. Basically I don't want to view the
'lines of code' that got me to my particular specific as a kind of 'black
box' that worked, but rather I want to understand the step by step and be
able to reprise whether the result arrived at was the specific I intended,
or that my originally imagined specific was appropriate to the analysis I
hoped to perform.
I start with a couple of assumptions: many researchers have had the same
problem before, gotten stuck, and asked for help (and probably gotten help
or gotten some suggestions). So I start with a couple of google searches:
cran r (avian) disease mapping vignettes
(some package) site:stat.ethz.ch (avian) disease mapping vignettes
and work the vignettes (as in type it out) to get a sense of how someone
else solved the problem. And maybe that's sufficient. But it allows me the
opportunity to examine a range of items (how is my ground cover data
structured, how are my disease instances structured, does the data I have
fit with the assumptions of a package I've elected to use, is there a
direct or indirect coersion path between packages that allows me to use the
tools of a given package upon an object of another & etc).
Then I can drill in and see if I can address my specifics and derive
results that are reasonable and understandable to me (and others).
Finally I read as much code as I can and see if I can get an inkling of
what is going on where and when, and potentially why.
http://stackoverflow.com/questions/19226816/how-can-i-view-the-source-code-for-a-function
And read source code where ever I can find it.
It is something of a grind to step back and do this, but it has given me a
better appreciation of the wealth of tools available. And even as they may
not be applicable to my problem at hand, to the extent that I understand my
problem, knowing about them informs me for analyzing future problems to
which they may be more properly applicable.
The foregoing process may not be applicable to your present needs, but I
imagine it may help with developing a sense of fluency within this system
of tools. Of course eventually I may regret every word here, but this
approach has helped me so far.
HTH,
Cheers
Chris
On Wed, Jun 10, 2015 at 3:50 PM, Joshua Onyango via R-sig-Geo <
r-sig-geo at r-project.org> wrote:
> Hello
> I am just started learning how to use R and do find this forum helpful.
>
> I want to carry out spatial statistics and disease mapping to explain
> possible differences in distribution in 8 regions of England.
> 1. How can I go about importing a nice map of England with eight regions
> plus boundaries similar to the one found on goggle -
> http://www.britaingallery.com/england_regions.php
> 2. Map cases per regions
> 3. Apply spatial statistics if possible to show how factors of interests
> such as latitude, mean temperature, precipitation, flock size and
> vegetation (thistles) could be used to explain differences is distribution
> of cases.
> Any response will be highly appreciated
> Thanks
> Josh
> [[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>
[[alternative HTML version deleted]]
More information about the R-sig-Geo
mailing list