Digitally Mapping the Republic of Letters –

Mapping Data

Digitally Mapping the Republic of Letters – This is a fascinating way to map social and political connections both spatially and temporally. (It’s also a testimonial to the nineteenth century postal service!) I find the emerging digitization of historic data (especially in the humanities) exciting and would love to see this similarly in the history of science.  How geographically clustered were schools of thought on ecology and evolution? Was there really an East/ West Coast divide in the early 20th century? Some would look to the mentoring system as a metaphor for familial lineage and influence but active correspondence would likely be more revealing and significant.

Thanks Jeni for pointing this article out!

And Happy GIS Day!

Wikileaks: Every Iraqi Death Mapped

Mapping Data

This map is powerful in its stark facts. It’s a simple point map on a Google base map but every point is report of one or more deaths, either armed forces, civilian or enemy, attributed to enemy fire, criminal activity, IED hazard, friendly fire, etc.  These categorizations raise more questions than they address, of course, but the toll on civilians is plain when you navigate the deaths. Major thoroughfares and intersections seem particularly dangerous to armed forces and civilians alike.

“The art of modelling range-shifting species”

Species Distribution Modeling

A podcast with Jane Elith, Michael Kearney and Steve Phillips in – Methods in Ecology and Evolution. They model the invasive Cane Toad to highlight the problems with purely correlative modeling when applying to novel climate scenarios from which it was trained (e.g. extrapolating to novel geography). In comparative tests of four methods with individual treatments, they demonstrate integrating mechanistic modeling and controlling the fit of models to increase reliability of models especially with species in non-equilibrium in novel circumstances. [link to online article]


Species Distribution Modeling, Tutorials

I can’t believe I missed this: World Statistics Day! Who knew there was a day to celebrate (if that’s the right verb for you) statistics?

So with that excuse, I’m reposting links to learning and using R from the R bloggers aggregate site which did commemorate World Statistics Day with a nice set of links to Tutorials for R.

On the subject of R, specific to Species Distribution Modeling, Robert Hijmans has a R package called dismo, which allows you to port your entire workflow to R,  from GBIF queries and georeferencing to modeling with Bioclim, Domain, Mahalanobis, GLM, GAM, Maxent, BRT, RF,  or SVMs.

More tutorial material on R and modeling in the future.