Point KDE and LSCV error

[Cleo Ferreira]

Hi Stefan

I'm trying to generate home ranges with LSCV and point KDE but when running the function, I get the error 'no point dataset'. 

I imported a .csv data and all my points are 'single points' and not 'multi-point'. 

There are also less than a 100 data points in my set. 

I also have the newest version installed but am still running into this problem. 

I know other users have had the same issue, have you maybe found any other solutions to this? 

Thank you so much,

Cleo

[Stefan Steiniger]

Hi Cleo,

sounds weird to me, like perhaps another layer is selected, instead of the one you want to process.

If you want, you can send me the dataset in private to sstei...@uc.cl and I will check what geometry type it may be - or if just one point is a little different (I would do this using the Attribute/Table View : a small icon indicates the type here - if it's just 100 points). 

best,

stefan

--
You received this message because you are subscribed to the Google Groups "OpenJUMP HoRAE Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ojhorae-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ojhorae-users/4554641c-4612-4e3f-9620-c0163819d280n%40googlegroups.com.

No sienta la obligación de contestar este mail fuera de horario laboral.

[Stefan Steiniger]

copying solution to list:

Hi Cleo,

So from what I see, the problem may be  that the data is in the units of degrees for lat/long - but the analysis is working with metric data - i.e. in meters. Hence, you would need to convert the data to a local cartographic system. I did this for instance with QGIS (in OpenJUMP it may be possible too) - converting the points to the UTM Zone 36S (see attached shapefile). Just for testing I did load the shapefile into OpenJUMP and added a weight attribute with all points having a weight of 1 (see also the attached OpenJUMP jml data file). Then I used the point KDE with h=1400, and created the contour for p=0.95%  - see attached screenshot. So, the data seem to be ok, when converted.

hope this helps,

Stefan