Community ecology is a sequence of decisions — how to transform the abundance table, how to measure diversity, whether groups differ, and which species drive the difference. Magic Stat puts that sequence in one dialog, from the transformations through to the indicator species, with figures ready for the manuscript.
The module follows the way an ecologist actually works through a community matrix: transform it (Hellinger, chord, Wisconsin, presence/absence, logs and roots), summarise it (alpha diversity, accumulation and rarefaction curves), compare groups (PERMANOVA, ANOSIM, PERMDISP, Mantel), ordinate it (PCA, PCoA, NMDS, CA, DCA, RDA, CCA), and go further with functional diversity, phylogenetic diversity, indicator species, TWINSPAN and variation partitioning.
The reference implementations for most of these live in vegan and its companions in R. If your lab already works in vegan, the honest advice is to stay there — the indices are the same, and vegan has more of them. Magic Stat matters when you want the analysis, the figures and the methods paragraph to come out of the same spreadsheet without a script, and when the person doing the analysis is not the person who wrote the script.
The module stops at the community matrix — the species and the sites. Everything downstream of that it covers: how the matrix is transformed, how it is summarised and compared, how it is ordinated, and how the result is written up. If what you need falls outside that boundary, the module will not pretend otherwise.
Dissimilarities: Bray–Curtis · Jaccard · Sørensen · Horn–Morisita · Euclidean · Manhattan · Canberra. Ordination methods: PCA · PCoA · NMDS · CA · DCA · RDA · CCA.
Permutations are set in the dialog for the permutation-based tests. With a grouping column, alpha indices are also compared between groups (ANOVA or Kruskal–Wallis per index, with FDR correction).
No. vegan in R is the reference for distance-based ecology and has more indices and options. Magic Stat covers the common workflow end to end from a spreadsheet and writes the report; if your analyses already live in vegan scripts, keep them.
Bray–Curtis is the convention for abundance data, Jaccard or Sørensen for presence/absence, and Euclidean for standardised continuous variables. Whatever you choose for the ordination, use the same one for PERMANOVA so the figure and the test describe the same thing.
Rarefaction rescales every sample to a common number of individuals, so sites with different sampling effort can be compared; accumulation shows how observed richness grows as samples or individuals are added in order. Both are computed in the accumulation tab, for richness and for the other indices.
Yes — figures export as PNG, SVG, PDF, JPEG or TIFF with the legends outside the plot and scientific palettes, and the analysis is written into a report with its settings.
No. Everything runs on your computer; the only automatic signal is an anonymous installation counter.