[Qgis-user] How to Identify the best subdataset of a dataset of points which follow a canonical distribution ?

[Lionel Roubeyrie]

Hello,
in first approch you can compute the general mean of your observations
and select only stations with rain mesureaments failing inside a
determinated tolerance around that mean. Another common use would be
to randomly remove some stations and look at the resulting variance
against the general variance (with all stations) and stop when a
tolerance is reached. The best approach would be to play with kriging
your observations and look at the resulting map when some stations are
removed.
Cheers

2011/1/25 Albin Blaschka <albin.blaschka at standortsanalyse.net>:
>
> Hello,
>
> another idea, unfortunately no solution, would be to ask at the r-sig-geo
> Mailinglist[1] - a list on the statistic-package R, which can be used
> directly in QGIS via the ManageR-Plugin...if you know a little bit of
> statistics and R... but could also be a little bit of overdoing if you have
> never worked with R before...
>
> Maybe an idea...
> Albin
>
> [1]
>
> R-sig-Geo at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>
>
> Am 25.01.2011 12:52, schrieb Anita Graser:
>
> Hi Nikos,
>
> I think this would be a perfect question for gis.stackexchange.com
> A manual approach (135 stations is not that much) would be to decide on a
> maximum distance between stations (depends on your analysis requirements I
> guess) and then buffer the station points accordingly. You can then manually
> remove points with redundant/overlapping geographic coverage.
> But I'm sure there are better solutions :)
> Best wishes,
> Anita
> On Tue, Jan 25, 2011 at 12:09 PM, nikos <vesnikos at gmail.com> wrote:
>>
>> Hello List,
>>
>>
>> I have an interesting problem, and I'd really appreciate if you could
>> give a couple insightful tips on how you'd solve the following:
>>
>> I want to order some meteo (rain mm/month for PSDI) data from my
>> corresponding national agency, but in their infinitive wisdom they
>> choose to place their weather stations in a pattern that follows
>> anything but a canonical distribution.
>>
>> I know this because they sent me a shp file with all their 136 stations
>> which are available to the general public - for a fee of couse.
>>
>> Now my problems arise on how I choose  which possible combination of my
>> dataset of stations correspond to a canonical distribution of the
>> mainland - keeping the set of the stations to a minimum as with each
>> station the cost goes up?
>> I want to be as much efficient one can be using GIS technology ;)
>>
>>
>> My initial thinking is to find a subdataset which have their voronoi
>> polygons created with the same area. But im not so sure if thats the
>> correct approach
>>
>> Any tips on how to solve this problem is greatly appreciated!
>>
>>
>>
>>
>> Ves Nikos
>>
>> _______________________________________________
>> Qgis-user mailing list
>> Qgis-user at lists.osgeo.org
>> http://lists.osgeo.org/mailman/listinfo/qgis-user
>
>
> _______________________________________________
> Qgis-user mailing list
> Qgis-user at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/qgis-user
>
>
> _______________________________________________
> Qgis-user mailing list
> Qgis-user at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/qgis-user
>
>



-- 
Lionel Roubeyrie
lionel.roubeyrie at gmail.com
http://youarealegend.blogspot.com