Dr. Sarah Goslee
@numberwright.bsky.social
Landscape ecology, multivariate statistics, geospatial everything. All opinions my own; none are official positions of my employer.
The book contains worked examples by the authors of many of the main ecology packages, and there’s an accompanying GitHub repo.
github.com/RCodingForEc...
I hope those of you who are my kind of nerds find something useful and interesting there.
github.com/RCodingForEc...
I hope those of you who are my kind of nerds find something useful and interesting there.
January 11, 2026 at 5:54 PM
The book contains worked examples by the authors of many of the main ecology packages, and there’s an accompanying GitHub repo.
github.com/RCodingForEc...
I hope those of you who are my kind of nerds find something useful and interesting there.
github.com/RCodingForEc...
I hope those of you who are my kind of nerds find something useful and interesting there.
Agreed, media coverage in this space lacks nuance. iNat has used machine learning and computer vision for a long time: that’s how their suggested identifications work. That’s a good use! What they’re proposing now is a generative AI partnership with Google. So many potential problems.
June 11, 2025 at 1:39 PM
Agreed, media coverage in this space lacks nuance. iNat has used machine learning and computer vision for a long time: that’s how their suggested identifications work. That’s a good use! What they’re proposing now is a generative AI partnership with Google. So many potential problems.
Image recognition is fine. There are lots of incredibly useful machine learning applications, including that kind of computer vision.
I object to using generative AI / LLMs for this (or, well, anything). Those models are trained on stolen data, environmentally expensive, and of dubious value.
I object to using generative AI / LLMs for this (or, well, anything). Those models are trained on stolen data, environmentally expensive, and of dubious value.
June 11, 2025 at 12:46 PM
Image recognition is fine. There are lots of incredibly useful machine learning applications, including that kind of computer vision.
I object to using generative AI / LLMs for this (or, well, anything). Those models are trained on stolen data, environmentally expensive, and of dubious value.
I object to using generative AI / LLMs for this (or, well, anything). Those models are trained on stolen data, environmentally expensive, and of dubious value.
800 in ARS (over 10%), and 1000 from NRCS.
February 15, 2025 at 2:25 PM
800 in ARS (over 10%), and 1000 from NRCS.
I am SO SECURE.
December 6, 2024 at 9:18 PM
I am SO SECURE.