Mark variograms: 3D(xyt) points + 1D marks

classic Classic list List threaded Threaded
6 messages Options
Reply | Threaded
Open this post in threaded view
|

Mark variograms: 3D(xyt) points + 1D marks

Tim Pollington
Dear all,

Are there packages that produce empirical *mark* variograms of 3D data (xyt) with 1D marks? I'm aware of the useful mvstpp package for creating spatial and temporal mark variograms but it doesn't completely fit my data structure.

Thanks, Tim.

        [[alternative HTML version deleted]]

_______________________________________________
R-sig-Geo mailing list
[hidden email]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo
Reply | Threaded
Open this post in threaded view
|

Re: Mark variograms: 3D(xyt) points + 1D marks

Tim Pollington
On reflection please ignore my question. One can't construct an xyt variogram with additional marks as one can't define a proper distance in an xyt space as different units are involved.

        [[alternative HTML version deleted]]

_______________________________________________
R-sig-Geo mailing list
[hidden email]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo
Reply | Threaded
Open this post in threaded view
|

Re: Mark variograms: 3D(xyt) points + 1D marks

edzer
No, but you can model them as a function of space and time, see e.g.
https://journal.r-project.org/archive/2016/RJ-2016-014/index.html

On 5/10/19 12:53 PM, Tim Pollington wrote:
> On reflection please ignore my question. One can't construct an xyt variogram with additional marks as one can't define a proper distance in an xyt space as different units are involved.
>
> [[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-Geo mailing list
> [hidden email]
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>

--
Edzer Pebesma
Institute for Geoinformatics
Heisenbergstrasse 2, 48151 Muenster, Germany
Phone: +49 251 8333081

_______________________________________________
R-sig-Geo mailing list
[hidden email]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo

pEpkey.asc (2K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Mark variograms: 3D(xyt) points + 1D marks

Tim Pollington
Thank you Edzer for your suggestion however I'm trying to avoid geostatistical methods because my data is disease cases recorded at household locations. Therefore strictly speaking there is not a random field that pervades the space between the households and so I must treat it as a spatiotemporal marked point process instead. I attemping to extend Stoyan et al's mark variograms for ST processes https://www.sciencedirect.com/science/article/pii/S2211675317300696 to my data.

Hopefully I'll have gotten a bit further by July so perhaps I can show you more if you're attending Spatial Statistics in Sitges.

Kind regards,

Tim.
________________________________
From: Edzer Pebesma <[hidden email]>
Sent: 10 May 2019 13:18
To: [hidden email]
Subject: Re: [R-sig-Geo] Mark variograms: 3D(xyt) points + 1D marks

No, but you can model them as a function of space and time, see e.g.
https://journal.r-project.org/archive/2016/RJ-2016-014/index.html

On 5/10/19 12:53 PM, Tim Pollington wrote:
> On reflection please ignore my question. One can't construct an xyt variogram with additional marks as one can't define a proper distance in an xyt space as different units are involved.
>
>        [[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-Geo mailing list
> [hidden email]
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>

--
Edzer Pebesma
Institute for Geoinformatics
Heisenbergstrasse 2, 48151 Muenster, Germany
Phone: +49 251 8333081

        [[alternative HTML version deleted]]

_______________________________________________
R-sig-Geo mailing list
[hidden email]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo
Reply | Threaded
Open this post in threaded view
|

Re: Mark variograms: 3D(xyt) points + 1D marks

edzer


On 5/11/19 6:11 PM, Tim Pollington wrote:
> Thank you Edzer for your suggestion however I'm trying to avoid
> geostatistical methods because my data is disease cases recorded at
> household locations. Therefore strictly speaking there is not a random
> field that pervades the space between the households and so I must treat
> it as a spatiotemporal marked point process instead. I attemping to
> extend Stoyan et al's mark variograms for ST
> processes https://www.sciencedirect.com/science/article/pii/S2211675317300696 to
> my data. 

Regrettably (an ironically, as an associate editor) I have no longer
access to that journal. Did you contact the authors of the paper whether
they have software they're willing to share?

>
> Hopefully I'll have gotten a bit further by July so perhaps I can show
> you more if you're attending Spatial Statistics in Sitges.

Looking forward to that!

>
> Kind regards,
>
> Tim.
> ------------------------------------------------------------------------
> *From:* Edzer Pebesma <[hidden email]>
> *Sent:* 10 May 2019 13:18
> *To:* [hidden email]
> *Subject:* Re: [R-sig-Geo] Mark variograms: 3D(xyt) points + 1D marks
>  
> No, but you can model them as a function of space and time, see e.g.
> https://journal.r-project.org/archive/2016/RJ-2016-014/index.html
>
> On 5/10/19 12:53 PM, Tim Pollington wrote:
>> On reflection please ignore my question. One can't construct an xyt variogram with additional marks as one can't define a proper distance in an xyt space as different units are involved.
>>
>>        [[alternative HTML version deleted]]
>>
>> _______________________________________________
>> R-sig-Geo mailing list
>> [hidden email]
>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>>
>
> --
> Edzer Pebesma
> Institute for Geoinformatics
> Heisenbergstrasse 2, 48151 Muenster, Germany
> Phone: +49 251 8333081
--
Edzer Pebesma
Institute for Geoinformatics
Heisenbergstrasse 2, 48151 Muenster, Germany
Phone: +49 251 8333081

_______________________________________________
R-sig-Geo mailing list
[hidden email]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo

pEpkey.asc (2K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Mark variograms: 3D(xyt) points + 1D marks

Tim Pollington
They have made spatial and temporal mark variograms (xy space with time as a mark, or 1D time with xy 2D marks, respectively) available via their mvstpp package.
The current problem is that I've just got many cases(points) who share the same location because they come from the same household. That's not so bad as marked pont processes can have coincident space coordinates, just not coincident spacetime coordinates.

However additionally I want to give a time mark not just to their onset but all the time they are infectious for until treatment (since infectivity probably rises during the infection). This would make the xyt space far too crowded. I want to avoid spatial jittering as I fear this will pollute the microscale semivariance which is of interest to me. The way I am trying is aggregating individuals into household units and introducing a fourth dimension C_i to represent total numbers of active cases at that time in hhld_i. My goal is a space-time variogram style plot but as a space-time marked variogram instead: by producing a spatial marked variograms stratified by time lag.

Sent from my mobile

________________________________
From: Edzer Pebesma <[hidden email]>
Sent: Sunday, May 12, 2019 9:04:39 PM
To: Tim Pollington; [hidden email]
Subject: Re: [R-sig-Geo] Mark variograms: 3D(xyt) points + 1D marks



On 5/11/19 6:11 PM, Tim Pollington wrote:
> Thank you Edzer for your suggestion however I'm trying to avoid
> geostatistical methods because my data is disease cases recorded at
> household locations. Therefore strictly speaking there is not a random
> field that pervades the space between the households and so I must treat
> it as a spatiotemporal marked point process instead. I attemping to
> extend Stoyan et al's mark variograms for ST
> processes https://www.sciencedirect.com/science/article/pii/S2211675317300696 to
> my data.

Regrettably (an ironically, as an associate editor) I have no longer
access to that journal. Did you contact the authors of the paper whether
they have software they're willing to share?

>
> Hopefully I'll have gotten a bit further by July so perhaps I can show
> you more if you're attending Spatial Statistics in Sitges.

[[elided Hotmail spam]]

>
> Kind regards,
>
> Tim.
> ------------------------------------------------------------------------
> *From:* Edzer Pebesma <[hidden email]>
> *Sent:* 10 May 2019 13:18
> *To:* [hidden email]
> *Subject:* Re: [R-sig-Geo] Mark variograms: 3D(xyt) points + 1D marks
>
> No, but you can model them as a function of space and time, see e.g.
> https://journal.r-project.org/archive/2016/RJ-2016-014/index.html
>
> On 5/10/19 12:53 PM, Tim Pollington wrote:
>> On reflection please ignore my question. One can't construct an xyt variogram with additional marks as one can't define a proper distance in an xyt space as different units are involved.
>>
>>        [[alternative HTML version deleted]]
>>
>> _______________________________________________
>> R-sig-Geo mailing list
>> [hidden email]
>> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>>
>
> --
> Edzer Pebesma
> Institute for Geoinformatics
> Heisenbergstrasse 2, 48151 Muenster, Germany
> Phone: +49 251 8333081

--
Edzer Pebesma
Institute for Geoinformatics
Heisenbergstrasse 2, 48151 Muenster, Germany
Phone: +49 251 8333081

        [[alternative HTML version deleted]]

_______________________________________________
R-sig-Geo mailing list
[hidden email]
https://stat.ethz.ch/mailman/listinfo/r-sig-geo