Francisco Rodriguez-Sanchez
LightNews LightNews
banner
frodsan.bsky.social
Francisco Rodriguez-Sanchez
@frodsan.bsky.social
3.3K followers 150 following 190 posts
Computational Ecologist. Researcher @unisevilla.bsky.social. ecology, biogeography, statistics, rstats, GIS, science.

https://frodriguezsanchez.net
Posts Replies Media Videos
frodsan.bsky.social
Two years ago we got together in a beautiful mountain to learn and develop #rstats packages for ecology

After {labeleR} (ecologyr.github.io/labeleR/) and {BlueCarbon} (ecologyr.github.io/BlueCarbon/), {mappestRisk} is now officially out! See thread by lead author @dario-ssm.bsky.social 👇
Collaborative development of R packages for ecology https://ecologyr.github.io/workshop/
November 20, 2025 at 7:17 PM
frodsan.bsky.social
The next @peercommunityin.bsky.social seminar by @simine.com looks interesting

peercommunityin.org/pci-webinar-...
Recognizing and responding to a replication crisis: Lessons from Psychology The last 15 years have been a turbulent period for the field of psychology. Failures to replicate, fraud, and evidence of questionable research practices have all contributed to a crisis of credibility. What has the field done in response to this crisis, and what lessons can be learned from psychology’s crisis? We often hear the self-correcting mechanisms in science invoked as a reason to trust science, but it is not clear what these mechanisms are, or how they could have prevented psychology’s crisis. Some quality control mechanisms, such as peer review for journals, or vetting for textbooks or for public dissemination, did not to provide much of a safeguard against invalid claims in psychology. Instead, I argue that we should look for visible signs of a scientific community’s commitment to self-correction. These signs include transparency in the research and peer review process, investment in error detection and quality control, and an emphasis on calibration rather than popularization. To regain or bolster trust, scientific fields should invest in these hallmarks of credibility.
November 5, 2025 at 11:48 AM
frodsan.bsky.social
Thank you @hormiga.bsky.social for writing this book. Beyond learning what may be right/wrong with current practices and many new tips, it makes you think throughout what kind of teacher you want to be and how to get there. Highly recommend

press.uchicago.edu/ucp/books/bo... #teaching #AcademicSky
Book cover of the Chicago guide to college science teaching
November 2, 2025 at 10:11 AM
frodsan.bsky.social
Still fascinated with #webR and its potential for teaching stats

Here students can play (even from their phones) and find out themselves how the p-value depends on sample size

pakillo.github.io/LM-GLM-GLMM-... #rstats
A screenshot of a statistics tutorial made with R, quarto and webR
October 22, 2025 at 6:36 AM
frodsan.bsky.social
Finally using {pins} #rstats package to share heavy files that can't be shared among collaborators through GitHub, and it works like a breeze! Supports data versioning, cache, etc. Thanks @posit.co

Here using Google Drive to share files, but can use many different servers, see pins.rstudio.com
library("pins") # specify folder in Google Drive to store objects board <- board_gdrive("https://drive.google.com/drive/folders/FOLDER_ID") # upload file board |> pin_write(big_dataset, "big_dataset.csv", type = "csv") # retrieve file (e.g. in another computer) board |> pin_read("big_dataset.csv")
October 2, 2025 at 12:27 PM
frodsan.bsky.social
Here's an example of the master document approach. Just move or silence chunks at will. Easier than copy-pasting full contents IMO
September 10, 2025 at 9:32 AM
frodsan.bsky.social
{grateful} 0.3.0 is now on CRAN. Hope it makes citing #rstats packages even easier!

pakillo.github.io/grateful/

New features 👇
Now possible to scan and cite dependencies from a package DESCRIPTION file (#62). Use pkgs = c("Depends", "Imports", "Suggests", "LinkingTo") or combinations of them to choose which dependencies to cite. Now possible to scan and cite dependencies from a single R script, ‘Rmarkdown’ or ‘Quarto’ document #65. Just provide the path to the file to the pkgs argument. Now possible to customise the language of citation paragraphs (#55), thanks to new arguments to cite_packages: text.start, text.pks and text.RStudio. New logical argument skip.missing allows the user to skip missing packages (those used somewhere in the project but not currently installed) #54. Slow down requests when querying dependencies for a large number of packages #61.
September 9, 2025 at 6:16 PM
frodsan.bsky.social
An extension of joint SDM to jointly model species abundances and intraspecific trait variation. #ecopubs @esajournals.bsky.social

doi.org/10.1002/ecy....
Joint species-trait distribution modeling: The role of intraspecific trait variation in community assembly The links between intraspecific trait variation and community assembly remain little studied, partially due to the lack of statistical methods to jointly model intraspecific trait variation and species abundances at the community level. Here, we extend the joint species distribution modeling (JSDM) framework into the joint species-trait distribution modeling (JSTDM) framework to explicitly link species abundances to phenotypic variation in traits for multiple species simultaneously. Using a case study of 65 tundra plant species abundances and 3 key functional traits measured across 325 sites, we show how the JSTDM approach (1) estimates the statistical associations among species abundances, species-level traits, and site-level traits, relative to environmental variation; (2) improves predictions on trait variation by using information on species abundances; and (3) generates hypotheses about trait-driven community assembly mechanisms. The JSTDM methodology presented in this study allows assessing the interplay between species abundances and traits at the community level, providing the much needed modeling tools to quantify the role of phenotypic trait variation in eco-evolutionary community assembly.
September 8, 2025 at 11:20 AM
frodsan.bsky.social
Sample size considerations for species co-occurrence models #ecopubs @esajournals.bsky.social

'While occupancy patterns are often robust to limited sample size, reliable inference about co-occurrence demands substantially larger datasets than many studies currently achieve'

doi.org/10.1002/ecy....
In the simplest scenario, there is high bias in the interaction parameter (used for co-occurrence inference) with less than 100 sites at high detection, and 400–1000 sites at low detection, depending on interaction strength. Strong co-occurrence is detected consistently above 200 sites with high detection probabilities, but weak co-occurrence is never consistently detected even with 2980 sites. We demonstrate that the mean predictive ability of the co-occurrence model is less affected by sample size
August 27, 2025 at 10:30 AM
frodsan.bsky.social
3. Now possible to skip packages that are used somewhere in the project but are currently unavailable (not installed). Will get a warning though, as those packages will not be cited.
A table of packages generated with grateful::cite_packages() where some packages have been omitted using the new argument `skip.missing = TRUE`
August 20, 2025 at 9:08 PM
frodsan.bsky.social
2. Now possible to scan and cite dependencies based on a package DESCRIPTION file
A table of {grateful} package dependencies, created with grateful::cite_packages()
August 20, 2025 at 9:08 PM
frodsan.bsky.social
1. Now possible to customise the language of citation paragraphs
A package citation paragraph in Spanish created with grateful::cite_packages()
August 20, 2025 at 9:08 PM
frodsan.bsky.social
I've added a few new features to {grateful}, the package that facilitates citation of #rstats packages, and I'd love feedback before submitting to CRAN 🧵
grateful: Facilitate citation of R packages The goal of grateful is to make it very easy to cite R and the R packages used in any analyses, so that package authors receive their deserved credit. By calling a single function, grateful will scan the project for R packages used and generate a BibTeX file containing all citations for those packages. grateful can then generate a new document with citations in the desired output format (Word, PDF, LaTeX, HTML, Markdown). These references can be formatted for a specific journal, so that we can just paste them directly into our manuscript or report. Alternatively, we can use grateful directly within an Rmarkdown or Quarto document. In this case, a paragraph containing in-text citations of all used R packages will (optionally) be inserted into the Rmarkdown/Quarto document, and these packages will be included in the reference list when rendering.
August 20, 2025 at 9:08 PM
frodsan.bsky.social
Rethinking Ecological Measures Of Functional Diversity

'none of the widely used metrics satisfy the requirements. Critical flaws leave them blind to the very biodiversity loss they're intended to detect. There's urgent need to develop new generation of FD metrics'

arxiv.org/abs/2506.17839 #ecopubs
Understanding functional diversity, the range and variability of species' roles and actions within their communities, is key to predicting and preserving the functions that sustain both nature and human well-being. In this paper, we provide a comprehensive review of the literature on functional diversity measurement. We begin by consolidating essential criteria that effective measures of functional diversity should meet. We then evaluate fifteen widely used functional diversity metrics against these criteria and assess their performance across six synthetic ecosystem scenarios where optimal behavior is known. Surprisingly, our analysis reveals that none of the widely used metrics fully satisfy all the established requirements, and all fail in at least one ecosystem scenario. In particular, we find that almost all metrics flagrantly violate set monotonicity and distance monotonicity, requirements that adding a novel species should increase diversity, and that the magnitude of that increase should grow with trait dissimilarity. We also find that metrics fail to decline when rare, functionally extreme species are lost, and even increase when a perfectly redundant species is added. These critical flaws leave them blind to the very biodiversity loss that functional diversity measures are intended to detect. Our findings underscore the urgent need to develop a new generation of functional diversity metrics that more accurately reflect ecological realities.
July 5, 2025 at 12:58 PM
frodsan.bsky.social
ICYMI githistory.xyz is a great way to navigate #git commits and visualise how a file has been changing across commits

Just replace github.com with github.githistory.xyz in the URL and enjoy! #rstats
June 19, 2025 at 5:34 PM
frodsan.bsky.social
And if you want dig even deeper {flow} is your friend: moodymudskipper.github.io/flow #rstats

Not only shows dependencies among functions, but also what happens within functions
June 13, 2025 at 11:01 AM
frodsan.bsky.social
{pkgnet} is also very nice: uptake.github.io/pkgnet

Produces an interactive map of functions within a package and a table with coverage info etc, as well as a complete report on #rstats package dependencies
Interactive function map of an R package, generated with https://uptake.github.io/pkgnet
June 13, 2025 at 11:01 AM
frodsan.bsky.social
What's your favourite approach to map relationships among functions in a #rstats package?

I really like github.com/kcf-jackson/..., e.g. here is the function map for {grateful} package
Function map for the 'grateful' package: https://pakillo.github.io/grateful/
June 13, 2025 at 11:01 AM
frodsan.bsky.social
Eltonian niche modelling: applying joint hierarchical niche models to ecological #networks

doi.org/10.1111/ele.... #ecopubs
There is currently a dichotomy in the modelling of Grinnellian and Eltonian niches. Despite similar underlying data, Grinnellian niches are modelled with species-distribution models (SDMs), whereas Eltonian niches are modelled with ecological-network analysis, mainly because the sparsity of species-interaction data prevents the application of SDMs to Eltonian-niche modelling. Here, we propose to adapt recently developed joint species distribution models (JSDMs) to data on ecological networks, functional traits, and phylogenies to model species' Eltonian niches. JSDMs overcome sparsity and improve predictions for individual species by considering non-independent relationships among co-occurring species; this unique ability makes them particularly suited for sparse datasets such as ecological networks. Our Eltonian JSDMs reveal strong relationships between species' Eltonian niches and their functional traits and phylogeny. Moreover, we demonstrate that JSDMs can accurately predict the interactions of species for which no empirical interaction data are available, based solely on their functional traits. This facilitates prediction of new interactions in communities with altered composition, for example, following climate-change induced local extinctions or species introductions. The high interpretability of Eltonian JSDMs will provide unique insights into mechanisms underlying species interactions and the potential impacts of environmental changes and invasive species on species interactions in ecological communities.
June 13, 2025 at 9:45 AM
frodsan.bsky.social
There you go!
df <- data.frame( web = c("https://example.com/", "https://example.com/"), value = c(NA, NA) ) for (i in 1:nrow(df)) { if (is.na(df$value[i])) { browseURL(df$web[i]) df$value[i] <- menu(c("option A", "option B", "option C")) } }
May 4, 2025 at 6:48 PM
frodsan.bsky.social
Not a package, but this simple code could do:
df <- data.frame( web = c("https://example.com/", "https://example.com/"), value = c(NA, NA) ) for (i in 1:nrow(df)) { browseURL(df$web[i]) df$value[i] <- readline(prompt = "Please enter value: ") }
May 4, 2025 at 6:34 PM
frodsan.bsky.social
Yes. R is most-cited item in OpenAlex but apparently missing from most-cited list provided by Google Scholar (available as Suppl Mat in Nature).

This is what they say about R:
April 15, 2025 at 2:08 PM
frodsan.bsky.social
Muy buenas las infografías de @creaf.cat sobre los robles y pinos ibéricos

www.creaf.cat/es/articulos...
March 23, 2025 at 11:26 AM
frodsan.bsky.social
Very interesting #ecopubs on 'Predicting interaction frequency in plant-pollinator #networks'

'Interaction frequencies are very difficult to predict, and using traits we cannot really do better than simply using abundance information'

doi.org/10.32942/X2S...
Using the best-predicting modelling approach (randomForest), we indeed see a slight decay in predictive quality with plant and pollinator compositional distance. Temporal distance played little role, although predictions for one year ahead (or back) were better than across the season or across multiple years. The overall predictive power of our models was low (Spearman’s 𝜌 ≈ 0.4), suggesting a very noisy system. Also, the most important predictor was abundance, as revealed by a parameter-free benchmark model that only used the cross-product of abundances to predict interaction frequency. Trait and phylogenetic information did not substantially improve predictive performance beyond abundance-based predictions. Across all studies, we failed to confirm a substantial contribution of ecological characteristics to pollinator-flower interaction frequency. One reason why predictions were relatively poor is that sampling effort is not standardised, and thus networks differed substantially in the observed number of interactions, network size, and interaction density. Also the pooling of networks across space or across time may have diluted preferences in the data, reducing their explanatory value. Finally, the majority of species in each network are rare, and the interaction information they provide may be much less relevant that that of common species. At present, we conclude that the frequencies of interactions are very difficult to predict, and using traits we cannot really do better than simply using abundance information.
March 21, 2025 at 8:14 AM
frodsan.bsky.social
New #statistics #ecopubs @esajournals.bsky.social:

Predictive models are indeed useful for causal inference doi.org/10.1002/ecy....

Sounds like a response to 'Predictive models aren't for causal inference' doi.org/10.1111/ele....
The subject of investigating causation in ecology has been widely discussed in recent years, especially by advocates of a structural causal model (SCM) approach. Some of these advocates have criticized the use of predictive models and model selection for drawing inferences about causation. We argue that the comparison of model-based predictions with observations is a key step in hypothetico-deductive (H-D) science and remains a valid approach for assessing causation. We draw a distinction between two approaches to inference based on predictive modeling. The first approach is not guided by causal hypotheses and focuses on the relationship between a (typically) single response variable and a potentially large number of covariates. We agree that this approach does not yield useful inferences about causation and is primarily useful for hypothesis generation. The second approach follows a H-D framework and is guided by specific hypotheses about causal relationships. We believe that this has been, and continues to be, a useful approach to causal inference. Here, we first define different kinds of causation, arguing that a “probability-raisers-of-processes” definition is especially appropriate for many ecological systems. We outline different scientific “designs” for generating the observations used to investigate causation. We briefly outline some relevant components of the SCM and H-D approaches to investigating causation, emphasizing a H-D approach that focuses on modeling causal effects on vital rate (e.g., rates of survival, recruitment, local extinction, colonization) parameters underlying system dynamics. We consider criticisms of predictive modeling leveled by some SCM proponents and provide two example analyses of ecological systems that use predictive modeling and avoid these criticisms. We conclude that predictive models have been, and can continue to be, useful for providing inferences about causation.
March 13, 2025 at 2:12 PM