Geoffrey Legault
LightNews LightNews
banner
legault.bsky.social
Geoffrey Legault
@legault.bsky.social
84 followers 210 following 8 posts
Data Scientist at Getty Images. Ph.D. in Ecology from CU Boulder.

Google Scholar Profile: https://scholar.google.com/citations?user=1Nh8dD0AAAAJ&hl=en
Posts Replies Media Videos
Reposted by Geoffrey Legault
bbolker.bsky.social
I often grimace/sigh when someone says "I'm having trouble with fitting this model by frequentist optimization, maybe MCMC sampling will work better" ... most of the geometry problems (multimodality, scaling, numerical instability) that plague hill-climbing algos also cause problems for samplers ...
December 28, 2025 at 7:22 PM
Reposted by Geoffrey Legault
jdushoff.bsky.social
This is a really cool project, I was lucky to have the opportunity to participate.

news.mcmaster.ca/mcmaster-res...
McMaster research team digitizes more than 100 years of Canadian infectious disease data
The new, publicly accessible database contains more than a million infectious disease incidence counts dating back to 1903.
news.mcmaster.ca
December 18, 2025 at 5:41 PM
legault.bsky.social
Haven't seen it posted yet, so I'm giving this paper led by @phyloecology.bsky.social a boost.

It models pest expansion onto hosts as a biased random walk on a graph (here, a phylogenetic tree). There's lots of way this approach could be expanded - hopefully it inspires
doi.org/10.1098/rspb...
Pest host expansion as a scale-free stepwise process across the host phylogeny
Abstract. The present and future host ranges of pests have important implications for ecology, economics and health. Most multi-host pests have phylogeneti
doi.org
December 1, 2025 at 3:31 PM
Reposted by Geoffrey Legault
jbakcoleman.bsky.social
Great piece on the absurdity of brute force multiverse analyses.

www.pnas.org/doi/10.1073/...
Robustness is better assessed with a few thoughtful models than with billions of regressions | PNAS
Robustness is better assessed with a few thoughtful models than with billions of regressions
www.pnas.org
October 22, 2025 at 5:29 PM
Reposted by Geoffrey Legault
Ecologists’ endless quest for automatic inference
statmodeling.stat.columbia.edu/2025/05/28/e...
Ecologists’ endless quest for automatic inference | Statistical Modeling, Causal Inference, and Social Science
statmodeling.stat.columbia.edu
May 28, 2025 at 6:32 PM
Reposted by Geoffrey Legault
masatoyamamichi.bsky.social
Excited to attend the ESJ (Ecological Society of Japan) meeting in Sapporo! Our lab has six presentations there:
esj.ne.jp/meeting/abst...
esj.ne.jp/meeting/abst...
esj.ne.jp/meeting/abst...
esj.ne.jp/meeting/abst...
esj.ne.jp/meeting/abst...
esj.ne.jp/meeting/abst...
ESJ72 シンポジウム S09-5
esj.ne.jp
March 14, 2025 at 12:30 AM
Reposted by Geoffrey Legault
shikharabhat.bsky.social
amnat.org/an/newpapers...

Nice to see Tahirah Williams' fantastic layperson summary of our ( @vishuguttal.bsky.social and I) Synthesis work on demographic stochasticity and eco-evolutionary dynamics out on the @asn-amnat.bsky.social webpage!

#evolsky #evolution #mathsky #physky #philbio 🧪🧬🌏
How Randomness Can Flip Evolution: New Study Uncovers Surprising Role of Population Noise
<p>Read about &ldquo;Eco-Evolutionary Dynamics for Finite Populations and the Noise-Induced Reversal of Selection&rdquo; by Ananda Shikhara Bhat and Vishwesha Guttal (Jan&nbsp;2025)</p><br/>
amnat.org
February 14, 2025 at 11:49 PM
Reposted by Geoffrey Legault
cboettig.bsky.social
ESIIL synthesis center's Innovation Summit for 2025 is focusing on "Environmental Tipping Points and Transformation, an immersive, hands-on event where innovation meets collaboration." Come join us in Boulder! cu-esiil.github.io/Innovation-S... Apps due Jan 31.
ESIIL Innovation Summit 2025
Summit 2025
cu-esiil.github.io
January 14, 2025 at 6:18 PM
Reposted by Geoffrey Legault
betanalpha.bsky.social
Do I use hierarchical modeling?

Bro my models have so many hierarchies they could form the basis of an unjust society.
December 14, 2024 at 4:53 PM
legault.bsky.social
For those interested in applying simulation-based inference for their ecology/evolution problems, this is a nice and brief overview of the Python package "sbi", a great toolkit for such purposes. arxiv.org/abs/2411.17337
sbi reloaded: a toolkit for simulation-based inference workflows
Scientists and engineers use simulators to model empirically observed phenomena. However, tuning the parameters of a simulator to ensure its outputs match observed data presents a significant challeng...
arxiv.org
November 27, 2024 at 3:41 PM
legault.bsky.social
This new paper by Misha Kummel and David Vasseur on stochastic simulation algorithms in @oikosjournal.bsky.social is absolutely fantastic. Clear, precise, and insightful. A real treat to read. Check it out if you can: nsojournals.onlinelibrary.wiley.com/doi/full/10....
Recent advances in stochastic simulation algorithms create opportunities to study new ecological and eco‐evolutionary problems
Incorporating stochasticity into ecological modeling is vital for understanding the structure and function of ecological systems, but stochasticity has been challenging to consistently introduce into...
nsojournals.onlinelibrary.wiley.com
November 22, 2024 at 3:17 PM