Advancing Ecological Informatics: A Tool for Implementing the Abundance–Preference Approach in Assemblage Analysis

Although the Abundance-Preference Diagram (APD) was introduced four decades ago as an entropy-based framework for classifying species along habitat-preference gradients, its uptake by the ecological community has remained marginal. The cause is not theoretical but operational: no existing software (e.g., vegan, ade4, CANOCO, PC-ORD, BIOMAPPER) implements the Degree of Preference index (DP) or the APD, which forces ecologists either to rebuild the algorithm from scratch or to abandon the method altogether. This barrier mirrors a broader concern in ecoinformatics, where the gap between published methods and accessible implementations continues to slow the adoption of quantitative tools by field ecologists. This study presents EcoAPD (https://apd.biodiversity.ma), the first open web-based platform dedicated to the APD method. The application takes a standard species-by-habitat abundance matrix and returns the full diagnostic suite (DP values, exclusive species curve, hyperbola-based preference classes, and the APD plot itself), without any programming or local installation. Three design choices set it apart from the descendants of CANOCO and ade4: 1) a dual-computation architecture that generates the normalized and non-normalized APD versions simultaneously at project creation, removing the need for users to commit to a method a priori; 2) a hybrid file/relational-database persistence layer with hash-based recompute avoidance, enabling visualizations to be regenerated without rerunning the full pipeline; and 3) a validation engine tailored to the irregularities of ecological matrices. The platform is evaluated on two real datasets and a synthetic size sweep. Vectorizing the preference-index core yields a speedup rising from 3.5× at 10 species to 85.6× at 1000 species, while full-pipeline latency is ~510 ms on a 93-species × 37-habitats matrix and ~2.3 s on a 114-species × 171-habitats waterbird monitoring matrix, the residual cost being dominated by habitat count rather than species count. A threshold-sensitivity ablation finds the classification stable at the 0.3 and 0.75 boundaries and sensitive at 0.1 and 0.5. EcoAPD makes a four-decade-old method operationally available, with potential to revisit historical assemblage studies that relied on abundance data alone.

@article{Farouki2026,
  title     = {Advancing Ecological Informatics: A Tool for Implementing the Abundance–Preference Approach in Assemblage Analysis},
  journal   = {International Journal of Advanced Computer Science and Applications},
  volume    = {17},
  number    = {7},
  year      = {2026},
  publisher = {The Science and Information Organization},
  author    = {Mohamed El Farouki and Mohamed Dakki and Meryem Ennakri and Soumia Ziti},
  doi       = {10.14569/IJACSA.2026.0170747},
  url       = {https://doi.org/10.14569/IJACSA.2026.0170747}
}