Subscribe to this thread
Home - General / All posts - Kriging Interpolation Vs calculating Density
katy_rafi5 post(s)
#01-Jul-20 09:09

Hi all,

I want to calculate density for a point dataset I have and create a hotspots map just based on point location and not values. I could not find a density model in Manifold, but instead used Kriging interpolation and tried variations to get the same results I had in ArcGIS map viewer (yellow to red heat map below).

I tried a constant value of 10 in Z to make it consider points locations only but the result was ranging from 9.999 to 10 and was not showing any variation.

Then I used a count column from my point dataset (values ranged 1-15) for Z value. The results in blue (low) to purple (high) below are not showing the right presentation of values or the density.

I understand interpolation is a different model , but still not certain why the empty areas are shown as high value areas in purple.

I used just the basic and auto version of the model in Manifold.

Would you please advise if there is any alteration in the model that can promote the result or if there is any model for calculating density in Manifold.

Thanks,

ColinD


1,960 post(s)
#01-Jul-20 09:25

Your attachments didn't attach.

Kernel Density Estimation (KDE) might help

http://www.georeference.org/forum/t118611.6


Aussie Nature Shots

dchall8
760 post(s)
#01-Jul-20 19:49

What is the projection for the drawing with the points?

tjhb

9,320 post(s)
#01-Jul-20 22:19

A point density function is not built in.

You will have to write it.

If you need help with that, provide some sample data.

dchall8
760 post(s)
#02-Jul-20 18:09

Then if you want to make a heat map, you have to build the density function first? Are there other ways to make a heat map in 9?

katy_rafi5 post(s)
#03-Jul-20 12:52
katy_rafi5 post(s)
#03-Jul-20 12:54

Hi,

I have attached some sample data, and would be nice if you can try a density map. Let me know if you can help with this. Thanks

Attachments:
GR_2019.zip

katy_rafi5 post(s)
#03-Jul-20 10:40

Please see the image attached.

Attachments:
manifold interpolation.jpg

Dimitri

6,104 post(s)
#03-Jul-20 12:04

Interpolation and point density visualizations are two very different things.

If you would like help doing point density visualization, provide your data set so somebody could provide some examples.

One possible approach is very simple:

1. create a grid of areas that covers the region of interest where each grid square (or hexagon, if you want to be stylish) covers a reasonable number of data points in regions where they are more prevalent. (There's a nice hexagonal tiling routine that's been published on the forum.)

2. using overlays, count the number of points that occurs in each grid square and put that number in the grid square.

3. Color the grid squares by the number of points in each and you have a simple density map. For a prettier map, create contours using the number of points in each grid square as the "height", or do an interpolation on the grid squares using the count of points in each. That gives you a map of density.

katy_rafi5 post(s)
#03-Jul-20 12:50

Thanks, I will try grids.

Sloots

473 post(s)
online
#08-Jul-20 13:40

This is what your data looks like with the M8 KDE add-in.

This is an example of a radius of 500 meters, Gaussian function. Chris

Attachments:
kde.png


http://www.mppng.nl/manifold/pointlabeler

Manifold User Community Use Agreement Copyright (C) 2007-2019 Manifold Software Limited. All rights reserved.