LoginSign Up
← Back to all sparks
I

impIndicator

ANALYTICS
Velocity0.0

R package impIndicator by b-cubed-eu — release notes from GitHub.

github.com/b-cubed-eu/impIndicator ↗

Biodiversity impact indicators settle their vocabulary before 1.0

biodiversityinvasive-speciesoccurrence-cubesuncertaintygbifr-package
◆Current state
impIndicator computes indicators of alien-species impact from GBIF-style occurrence cubes, producing species-level, site-level and regional measures with visualisation. The latest release renames the three headline functions to compute_species_indicator(), compute_site_indicator() and compute_regional_indicator(), drops the division by total occupied sites, and fixes the exponential transformation of impact categories into scores. It is part of the b-cubed-eu family and leans on sibling tooling rather than reimplementing it.
◆Where it's heading
Two threads run through the recent releases. One is uncertainty: 0.6.0 wires in dubicube for cross-validation and uncertainty estimation on the indicators, moving output from point estimates toward quantified confidence. The other is scoping and naming — user-supplied sf regions in 0.4.0, occurrence-cube construction in 0.5.0, then the 0.6.1 rename — the pattern of a package tightening its public vocabulary as it approaches a stable release.
◆Prediction
With the naming settled and uncertainty estimation in place, the next step is most likely consolidation toward a 1.0 — documentation and vignettes against the renamed functions rather than further indicator types.

◆Recent moves

  1. 4mo ago

    Indicator functions renamed; scores no longer site-normalised

    All three entry-point functions are renamed for consistency and overall becomes regional, which breaks existing scripts. More consequential than the rename: species and regional indicators are no longer divided by the total number of occupied sites, and a bug in the exponential transformation from impact categories to scores is fixed — both change the numbers this package reports.

    View source ↗
  2. 5mo ago

    Uncertainty estimation for impact indicators via dubicube

    Overall impact indicators gain uncertainty calculation and visualisation through the dubicube package, with prepare_indicators_bootstrap() assembling the parameters for cross_validate_cube(). Delegating to a sibling b-cubed package rather than writing its own resampling keeps the statistical machinery in one place across the family.

    View source ↗
  3. 7mo ago

    Export impact_cube_data() for building impact occurrence cubes

    An exported function for creating an occurrence cube carrying impact data, which closes the gap between raw occurrence records and the cube the indicator functions expect. The Acacia cross-validation example in the README gives the uncertainty work from the following release something concrete to stand on.

    View source ↗
  4. 8mo ago

    Indicators can be computed for a user-supplied region

    The indicator functions intersect the cube with a supplied sf region and compute only over cells falling inside it, so the analysis extent is no longer dictated by whatever cells the cube happens to contain. This is the change that makes national or protected-area reporting practical.

    View source ↗
  5. 8mo ago

    impIndicator 0.3.2

    Documentation only: the example switches to a GBIF cube rather than building one from occurrence records. It nudges users toward the standard cube format but changes no behaviour.

    View source ↗
  6. 9mo ago

    impIndicator 0.3.1

    A version-number correction and nothing else — a housekeeping tag between substantive releases.

    View source ↗