| Author | Post |
|---|
brian.burke
Member
| Joined: | Fri Aug 21st, 2009 |
| Location: | |
| Posts: | 4 |
| Status: |
Offline
|
| Mana: |  |
|
Posted: Wed Jun 16th, 2010 04:30 pm |
|
Is there any way to calculate and/or plot results from spatial statistics? At the moment, I'd like to plot the Center of Gravity (CG) and Inertia (standard deviation of distance from CG in two directions), but I'd also like to calculate things such as Ripley's K function or similar statistics. See attached image for the CG and Inertia (red lines).
Thanks,
-BrianAttachment: CG.jpg (Downloaded 18 times)
|
Matt Wilson
Internal
| Joined: | Tue Aug 12th, 2008 |
| Location: | Hobart, Australia |
| Posts: | 66 |
| Status: |
Offline
|
| Mana: | |
|
Posted: Tue Jun 22nd, 2010 01:06 am |
|
Hi Brian,
At this stage Eonfusion doesn't contain very many operators for specific statistical operations. In the broad sense there are two possibilities for achieving these analyses and visualizations:
1) Use Eonfusion's built-in operators along with code inside an Evaluate Expression operator. The essential process is to: - Link the points somehow. Creating a surface from the points enables the use of a "for each vertex" loop within the Evaluate Expression operator. Another option is to use the Link Neighboring Vertices operator so that you can traverse the whole point set.
- Process the point set, applying the required analysis. This is achieved via the relevant code in the Evaluate Expression operator.
- Generate a new data set for visualization. In the CG/Inertia example this would require creation of two lines and a point, followed by migration of the analysis results onto the vertices of the lines and the point in order to locate them in space.
2) Write an Add-in operator using Eonfusion's API. This requires more than option (1) in the way of programming know-how but is a more direct way to achieve the result, because the API will allow more flexible access to the input point set as well as the ability to more easily encode the output results.
As a starting point, let me know whether either of these options are likely to provide a practical way ahead for you. If you'd like to pursue one of these then I'm more than happy to help out with the next steps. Another option is to hire one of our developers to write the relevant add-in operator(s) for you.
Best wishes,
Matt
|
brian.burke
Member
| Joined: | Fri Aug 21st, 2009 |
| Location: | |
| Posts: | 4 |
| Status: |
Offline
|
| Mana: | |
|
Posted: Tue Jun 22nd, 2010 06:29 pm |
|
Hi Matt, thanks for the reply.
I don't think the first option will work. Part of the calculation of the inertia lines (the part that determines the angle the lines should be at) uses principle components analysis, which I would not be able to code into Evaluate Expression.
The second option would probably work just fine, but I'm not inclined to start such a complex task at the moment. I've done all these analyses in R, so if there was an easy way to get R code to work as an add-in, that might be more attractive.
-Brian
|
Matt Wilson
Internal
| Joined: | Tue Aug 12th, 2008 |
| Location: | Hobart, Australia |
| Posts: | 66 |
| Status: |
Offline
|
| Mana: | |
|
Posted: Wed Jun 23rd, 2010 05:58 am |
|
Hi Brian,
The Eonfusion Add-in development API is sufficiently flexible to allow you to pass data to (and retrieve results from) R routines.
To get a general starting guide to the Add-in API, have a look at the following page in the manual:
http://www1.eonfusion.com/manual/index.php/Add-in_development
And have a look at this page for notes on how to integrate R with .NET:
http://learnserver.csd.univie.ac.at/rcomwiki/doku.php?id=wiki:using_r_d_com_and_rcom_in_.net_applications
For security reasons Eonfusion won't allow access to unmanaged code from an Add-in operator, unless the Add-in operator is specially signed. We can provide the necessary key upon request.
Please don't hesitate to ask if you have any questions.
Best wishes,
Matt
|
Current time is 06:21 pm | |
|