naturalis / trait-geo-diverse-dl

Ecological Niche Modelling Using Deep Learning

trait-geo-diverse-dl

This repository develops the application of deep learning to species distribution modelling (DL-SDM). The repository consists of three sub-projects:

• pilot - contains the same occurrences and GIS data as used in this application of a MaxEnt SDM to determine ecological niches of the world's ungulates. The purpose of this subproject is to assess the performance of DL-SDM in a like for like comparison with MaxEnt modelling.
• data_extended - incorporates additional occurrence data. The purpose of this project is to develop insight into the number of occurrences that needs to be collected for credible modelling.
• data_GIS_extended - also incorporates additional GIS data.

trait-geo-diverse-dl's Issues

cleanup

• use shorter file names
• do not use spaces in file names
• provide a README.md file in every directory describing the directory contents
• do not commit OS-specific helper files such as Thumbs.db
• use descriptive commit messages, not "Commit message"
• do not dump everything in one folder. For example, script is for scripts only
• make the ipynb files viewable, i.e. not too big
• use comma- (or tab-) separated files instead of Excel spreadsheets

import Laurens's prototype code

Would be great to have some code samples or prototype code that Laurens Hogeweg developed for his earlier proofs of principle here

load data

We will use the cleaned occurrence data sets from here: https://github.com/naturalis/trait-geo-diverse-ungulates/tree/master/data/filtered

Ideally we will ensure we stay synchronized, so perhaps that location can be a git module inside this one.