Q on New Environmental Layers / Data Management Needs for SDMs?
I’m taking a "Programming for GIS" class this summer, using Python.
We have a final project due about a month and I need to choose a topic.
I’d like to do a programming project related to the Maxent modeling in the Park.
Integrating the maxent.jar file into a Python script for ArcGIS has already been integrated into Desktop ArcGIS, as shown here:
I’d like to work on a script like this for the Smokies to use.
I’m wondering if you have any suggestions for programmatically accomplishing data management tasks related to the Maxent models?
For example, I remember you mentioning you’ve written some programs that generate basic metadata.
That sounds valuable, and worth incorporating into a Python script.
Can you share the Python script you wrote for generating metadata? I’d like to include metadata creation as a step in whatever script I come up with.
Another useful step in the script might be actually assigning X,Y Coordinates to input data using Python, sort of data "pre-processing" the coordinates coming out of the ATBI database for analysis in ArcGIS – and not to get too greedy, but I’m wondering if extracting point coordinates from Chuck Cooper’s MySQL database could be accomplished programatically.
As an aside, I’m also wondering if you’ve got new environmental layers? I can use the old ones I already have, but I might as well use the new ones if they’re available.
For one more potential programming application:
I understand that Todd Witcher is interested in getting his interns involved in doing field verification of the species distribution models.
I was thinking it might be possible to write a program that could choose random sampling locations within a given buffer zone (for example, random sample locations in areas defined as the union of reasonable access to a trail, and either above or below the threshold for predicted "presence" or "absence."
From an earlier project, I seem to recall that the threshold value of "predicted presence" versus "predicted absence" was different for each species modeled, so that might be difficult to accomplish programmatically, but it’s definitely an interesting research question.
If doable, I’m thinking this’d be a way to generate a lot of sample sites in a consistent, reproducible manner.
Any thoughts on that?
I know summer’s a busy time for you –
Would love to check in via phone if that works better.