[GRASS-dev] temporally aggregate a vector time series

GRASS GRASS developer
1

Hello devs

I’m wondering whether it is possible to aggregate space time vector data sets (STVDS).

I have a daily time series of vector maps (points) and I would like to aggregate it weekly. Is that possible somehow with t.vect.algebra? I blindly assumed there was something like t.rast.aggregate for vectors too, but there’s no… I think maybe it is possible with t.vect.algebra given that the operation would be equivalent to v.overlay operator=or, but as with t.rast.algebra, those manual pages are just a mistery to my limited understanding… Can anyone give any hints? or hall I discard vector time series approach completely?

Thanks a lot in advance!

Best,
Vero

2

Hi,
unfortunately, there is no vector aggregation available at the attribute level. However you can transform your time series of vector point data into coarse raster data time series, aggregate it and sample the result with the point coordinates (t.vect.observe.strds). Make sure that vector points belong to single pixels, hence one point per pixel.

Best regards
Soeren

2016-09-07 14:59 GMT+02:00 Veronica Andreo <veroandreo@gmail.com>:

Hello devs

I’m wondering whether it is possible to aggregate space time vector data sets (STVDS).

I have a daily time series of vector maps (points) and I would like to aggregate it weekly. Is that possible somehow with t.vect.algebra? I blindly assumed there was something like t.rast.aggregate for vectors too, but there’s no… I think maybe it is possible with t.vect.algebra given that the operation would be equivalent to v.overlay operator=or, but as with t.rast.algebra, those manual pages are just a mistery to my limited understanding… Can anyone give any hints? or hall I discard vector time series approach completely?

Thanks a lot in advance!

Best,
Vero

grass-dev mailing list
grass-dev@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-dev

3

Hello Sören,

Thanks for your answer :slight_smile:

2016-09-07 21:57 GMT+02:00 Sören Gebbert <soerengebbert@googlemail.com>:

Hi,
unfortunately, there is no vector aggregation available at the attribute
level.

Ok, I understand... attribute aggregation is the problem, not geometry

However you can transform your time series of vector point data into
coarse raster data time series, aggregate it and sample the result with the
point coordinates (t.vect.observe.strds). Make sure that vector points
belong to single pixels, hence one point per pixel.

Yes, after I wrote to the list, I thought of this by-pass, too
(v.to.rast, create the strds and register maps, to then aggregate as
desired and sample with t.vect.observe.strds). However, I'll loose the
attributes in the original vector points.
So, I thought of better splitting the original vector by weeks in R
(split), then write all weekly vector maps into grass (writeVECT) and
register the STVDS.

btw, is that paper regarding temporal algebra you mentioned once
already published?? I would like to study that.

Thanks again :slight_smile:

Best regards,
Vero

4

Hei Sören!

[snip]

I wish it was, but it is not. I have to spend much more time in writing it.
However, i have attached the AGU poster of Thomas Leppelt and me regarding
the algebra. So you can print it in poster format and put it on your wall
for reading. :slight_smile:

Well, I could do that! :smiley: :smiley: LOL!

Thanks much!
Vero