[R-sig-Geo] Classes for spatio-temporal data?

Edzer Pebesma edzer.pebesma at uni-muenster.de
Thu Sep 10 09:23:45 CEST 2009


Christian, I'm very interested in both: collaborating and watching.

Some previous work that I'm aware of: package trip (on CRAN) extends
SpatialPointsDataFrames basically in the attributes, by defining two
columns of which one is the time stamp and the other the (animal or
trip) ID. In chapter 6 of "Applied Spatial Data Analysis with R", this
example is discussed, and an example is given for creating a class for
space-time grids by extending SpatialGrid and SpatialGridDataFrame. The
example R code is found at http://asdar-book.org/book/csdacm.R

As with spatial data, you'll be facing the decision whether to make your
spatial dimension regular (discrete, "raster"), or continuous
("vector"). In your example, space is continuous ("Points"), and time is
regular. Will your methods of analysis and visualisation only be
suitable for regular time data, or can/should they handle continuous as
well?

Blair Christian wrote:
> Hi All,
>
> I'm starting to work with spatio temporal data, and wondered if there were
> any extensions to the sp package that include a temporal component. (for
> example, any classes that might contain 3 objects/data frames, one for the
> spatial information, one for the temporal information, and one for the
> data).
>
> My first goals are (1) getting some basic classes together for my own narrow
> set of circumstances (focus on SpatialPointsDataFrames with regular hourly
> time series), and (2) getting some basic visualization together.  So far
> I've used some kriging (in combination with the animation package) for
> making movies of kriging predictions along with some FDA of the temporal
> side (fourier series to examine the strong diurnal pattern).  It would be
> nice to be able to eventually make some visualizations in a way that could
> be shared with others, and I know that kml allows you to share both the raw
> data and use, say, google earth to visualize the time element.
>
> I know it sounds ambitious, but it's something I'll be working on off and on
> over the next year.  I have enough data at the moment that it's worth my
> while to add some structure ( (S4?) classes/methods) to my dusty decks.
>
> If anybody is interested in collaborating or watching, I'll try and put up a
> website at some point, maybe with a skeleton package.
>
> Cheers,
> Blair
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>   

-- 
Edzer Pebesma
Institute for Geoinformatics (ifgi), University of Münster
Weseler Straße 253, 48151 Münster, Germany. Phone: +49 251
8333081, Fax: +49 251 8339763 http://ifgi.uni-muenster.de/
http://www.springer.com/978-0-387-78170-9 e.pebesma at wwu.de



More information about the R-sig-Geo mailing list