Skip to content

Nuée 1.0

Choose a tag to compare

@4rtefakt 4rtefakt released this 15 Nov 22:02
55d07db

Maps and spatial statistics, community ecology, population monitoring with TRIM and distance sampling, random forests and boosted trees, a guide to the right analysis, and Nuée on macOS.

Install

Windows

  1. Download Nuee-Setup-1.0.0.exe below.
  2. Run it and choose to install for yourself (no administrator rights needed) or for everyone on the computer. It replaces an earlier version, and keeps its shortcuts and taskbar pin; from 0.9.1, a click in the status bar installs it.
  3. The installer isn't code-signed, so Windows may say it protected your PC: click More info, then Run anyway.

R is included; nothing else to install. Windows 10 or 11, 64-bit. Uninstall from Settings › Apps.

macOS (new)

  1. Download Nuee-1.0.0-arm64.dmg (Apple silicon, macOS 14 or later) or Nuee-1.0.0-x64.dmg (Intel, macOS 13 or later) below.
  2. Open it and drag Nuée into Applications.
  3. The application isn't signed with an Apple Developer ID, so the first time macOS says it couldn't check it: click Done, then Open Anyway in System Settings › Privacy & Security, and confirm.

R is included, as on Windows: nothing else to install, and an R already on the Mac is neither used nor changed.

Linux

  • Debian, Ubuntu and derivatives: download Nuee-1.0.0-amd64.deb and run sudo apt install ./Nuee-1.0.0-amd64.deb. apt brings R and the R packages Nuée needs.
  • Other distributions: use Nuee-1.0.0-x86_64.AppImage (needs libfuse2) or Nuee-1.0.0-x64.tar.gz (unpack, then ./nuee).
    • Install R and its packages first: install.packages(c("httpuv", "jsonlite", "openxlsx", "haven", "readxl", "foreign", "nlme", "mgcv", "MASS", "rpart", "cluster", "RSQLite", "ranger")).
    • See the installation page.
  • Older R: Nuée uses the computer's R. With an R older than 4.4 (Ubuntu 22.04 and 24.04, Debian 12), .rds and .RData files don't open, because reading one could run code (CVE-2024-27322).

The macOS and Linux packages are built and checked on GitHub from this release's tag, then added here.

New in 1.0

  • Nuée on macOS: Nuée.app with its own R, for Apple silicon and Intel Macs. Every engine check also runs on both kinds of Mac, with the same results as on Windows and Linux.
  • Maps and spatial data: a project's coordinates and their system (WGS 84, Lambert-93, UTM, LAEA Europe…), Plots › Map with an OpenStreetMap or IGN background only when you ask for one, GeoPackage layers in and out for QGIS, raster values at each site (elevation, land cover, climate) from GeoTIFF or ESRI ASCII grids, Moran's I with its correlogram and semivariogram, and a spatial smooth in GAMs. The same numbers as PROJ, terra, spdep, gstat and mgcv.
  • Community ecology: PCoA, NMDS, PERMANOVA with the dispersion of each group, Mantel tests, canonical correspondence analysis, indicator species, rarefaction and accumulation curves with richness estimators. The same numbers as vegan and indicspecies.
  • Population monitoring: TRIM trends and indices from count surveys, the same as rtrim; distance sampling from line and point transects, with or without covariates in the detection function, the same as the Distance package, with its few documented differences.
  • More models: one model per response with p values corrected over the species; zero-inflated Poisson and beta distributions in GLMs and GAMs; random forests (ranger) and boosted regression trees (as gbm computes them).
  • Which analysis? (Model › Which analysis?): your question in plain words, then the recommended analysis, why, its conditions and the alternatives, and New model already filled in. A regression's Summary now says in words whether the model's conditions hold.
  • Writing: the Results text, figures composed for a journal (vector PDF, SVG, PNG or TIFF at 300 or 600 dpi), reproducible Quarto documents that recompute every cited number from the raw data, and Zotero group libraries.
  • Data workbench: Find and replace, categories renamed, merged and reordered, a unit and a description for each variable, and import from KoboToolbox and ODK.
  • Export to R for every analysis type, with R alone (ranger for random forests).
  • Faster on large tables and spectra: a PCA 10 times faster, a PLS 8 times, permutation tests 14 times, with identical results.
  • English by default for new installations (View › Language); in French, numbers now have a decimal comma.
  • Fixed: a GLM or GAM with a single random effect and no curve lost that effect's variance; Predict a file failed after a Savitzky–Golay step; kept adjusted means were lost on reopening a project; each update removed Nuée's pin from the taskbar.

Projects saved by 1.0 still open in 0.9.9. The full list is in Help › What's new and CHANGELOG.md. Documentation and tutorials: 4rtefakt.github.io/nuee; website: nuee.io.

Cite

Boutin, C. (2026). Nuée (version 1.0.0) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.24263015

The DOI 10.5281/zenodo.24263014 stands for every version. Model › Text for the Methods section also writes a citation for you.

Checksum

SHA-256 of Nuee-Setup-1.0.0.exe:
d6127b53745e26a9eb753c264d22fc12f10e2ed9b48be137ab5d7c04459f91d1