A checklist for choosing between #rstats packages

The paradox of choice can at times be a challenge. There are well over 10,000 packages on CRAN now (likely 16,000), and there have been suggestions on how to find what you need but not necessarily on how to choose between alternatives. Here is a brief checklist that I used to contrast two similar packages for doing meta-analyses in R (summarized in a preprint qualitatively).

regularly maintained
recently updated
package maintainer on GitHub
manual available
vignettes available
used/published in similar projects
aligned workflow
semantics intuitive
contemporary grammar
functions that get the job done
arguments to support needs
visualization options (if needed)
dependencies reported/reasonable
connects to other packages


It is great to have different choices, and it is important to explore different alternatives because each can lead to potentially different adventures.

A vision statement describing goals for Ecology @ESAEcology #openscience

Many aspects of the journal Ecology are exceptional.  It is a society journal and that is important. The strength of research, depth of reporting, and scope of primary ecological research that informs and shapes fundamental theory has been profound.  None of these benefits need to change.  Nonetheless, research that supports the scientific process and engenders discovery can always evolve and must be fluent.  So must the process of scientific communication including publications through journals.  With collaborators and support from NCEAS and a large publishing company, I have participated in meta-science research examining needs and trends in the process of peer review for ecologists and evolutionary biologists, i.e. Behind the shroud: a survey of editors in ecology and evolution published in Frontiers in Ecology and the Environment or biases in peer review such as Systematic Variation in Reviewer Practice According to Country and Gender in the Field of Ecology and Evolution published in PLOS ONE.  In total, we have published 50 peer-reviewed publications describing a path forward for ecology and evolution in particular with respect to inclusivity, open science, and journal policy.  Ideally, we have identified at least three salient elements for journals relevant to authors, referees, and editors, and four pillars for a future for scholarly publishing more broadly.  The three elements for Ecology specifically would be speed, recognition, and more full and reproducible reporting.  The four pillars include an ecosystem of products, open access, open or better peer review, and recognition for participation in the process .


Goals to consider

  1. Rapid peer review with no more than 4 weeks total for first decision.
  2. A 50% editor-driven rejection rate of initial submissions.
  3. Two referees per submission if in agreement (little to no evidence more individuals are required).
  4. Double the 2017 impact factor to ~10 within 2 years and return to top 10 ranking in 160 of journals listed in field of ecology.
  5. Further diversify the contributions to address exploration, confirmation, replication, consolidation, & synthesis.
  6. Innovate content offering to encompass more elements of the scientific process including design, schemas, workflows, ideation tools, data models, ontologies, and challenges.
  7. Allow authors to report failure and bias in process and decision making for empirical contributions.
  8. Provide additional novel material from every publication as free content even when behind paywall.
  9. Develop a collaborative reward system for the editorial board that capitalizes on existing expertise and produces novel scientific content such as editorials, commentaries, and the reviews as outwardly facing products. Include and invite referees to participate in these ‘meta’ papers because reviews are a form of critical and valuable synthesis.
  10. Promote a vision of scientific synthesis in every publication in the Discussion section of reports. Request an effect size measure for reports to provide an anchor for future reuse (i.e. use the criteria proposed in ‘Will your paper be used in a meta‐analysis? Make the reach of your research broader and longer lasting’).
  11. Revise the data policy to require data deposition – at least in some form such as derived data – openly prior to final acceptance but not necessarily for initial submission.
  12. Request access to code and data for review process.
  13. Explore incentives for referees – this is a critical issue for many journals. Associate reviews with Publons or ORCID.
  14. Emulate the PeerJ model for badges and profiles for editors, authors, and
  15. Remove barriers for inclusivity of authors through double-blind review.
  16. Develop an affirmative action and equity statement for existing publications and submissions to promote diversity through elective declaration statements and policy changes.
  17. All editors must complete awareness training for implicit bias. Editors can also be considered for certification awarded by the ESA based on merit of reviewing such as volume, quality of reviews, and service. Recognition and social capital are important incentives.
  18. Develop an internship program for junior scientists to participate in the review and editorial process.
  19. Explore reproducibility through experimental design and workflow registration with the submission process.
  20. Remove cover letters as a requirement for submission.


I value our community and the social good that our collective research, publications, and scientific outcomes provide for society.  However, I am also confident that we can do more.  Journals and the peer review process can function to illuminate the scientific process and peer review including addressing issues associated with reproducibility in science and inclusivity.  Know better, do better.  It is time for scientific journals to evolve, and the journal Ecology can be a flagship for change that benefits humanity at large by informing evidence-based decision making and ecological literacy.


Ecological network flavors: many-to-many, few-to-many, and few-to-many spatially

Recent conference attendance inspired me to do a quick typology of networks that were presented in various talks. All were done in R using a few different packages.
All were interested in diversity patterns.
None were food webs.


many-to-many: many plant species and many pollinators for instance

few-to-many: mapping the associated set of pollinators to one flowering species

few-to-many: replicated mapping of diversity for one taxa to a single species of another either nested or spatially contrasted.


Network analyses are amazing. I need to learn more!

Can you also map interactions onto other interactions?



Sharing strategies for #ESA2016 #openscience #scicomm

Meetings are an excellent opportunity to not only communicate your science but secure feedback. I propose the more you give, the more get.


There are at least the following five open-science products associated with any contribution (presentation or poster) to share with your colleagues and a much wider online audience prior to the meeting.  [ green text = hyperlinks ]

Open-science products to share for a meeting

  1. The slide deck or poster can be published on SlideShare.
  2. Your data-science workflow, code, and EDA can be published as an r-markdown on GitHub.
  3. The primary or derived summary data (if you are not ready to go public yet) can be published on figshare (and/or included in GitHub repo).
  4. Most journals accept submissions that have been pre-printed. Consider sharing your draft paper on PeerJ or bioRxiv. Not at that stage? Do a blog post instead.
  5. Record a video abstract to share the main finding of your talk and post to YouTube or Vimeo. This could attract a larger audience to the conference presentation and does an incredibly useful service in communicating science to the public and others that do not attend the conference.


I enjoy the process of science way too much, and I am easily distracted.  How did I do in preparing for my ESA talk this year on microenvironmental change under desert shrubs?  I scored a total of 4 our of 5 . Feel free to click on bolded text below to see materials and provide feedback. Each is absolutely a work-in-progress like the experiment itself (we need at least one more year of data). However, I am hoping it is a good time to share ideas now and see if we can do better next year in the field.

Deck on SlideShare

Code on GitHub

Data on GitHub

Video abstract (went a bit crazy here and did two). Field and in office versions. Very high cheese factor in both (hard to be natural on camera).

Science is a process. Share your steps.