abrahamnunes/diversity

Diversity

Tools for measuring heterogeneity

Installation

pip install git+https://github.com/abrahamnunes/diversity.git

Documentation

Documentation is provided in the Jupyter notebooks located in the docs/ directory.

Measures implemented

• Berger-Parker Diversity index: berger_parker_diversity
• Berger-Parker Dominance index: berger_parker_dominance
• Rényi heterogeneity (aka. Hill numbers): renyi
  • Includes decomposition (renyi_decomp) into $\alpha$ (renyi_alpha) and $\beta$ components
• Rényi entropy: renyi_entropy
• Coefficient of variation: coef_variation
• Community weighted mean: func_cwm
• Freeman index: freeman_index
• Functional divergence: func_div
• Generalized Entropy Index: gei
• Gini coefficient (based on Lorenz curve): gini_lorenz
• Gini-Simpson index: gini_simpson
• Hartley heterogeneity: hartley
• Hartley entropy: hartley_entropy
• Lincoln index (base, Chapman, and Bayesian estimators): lincoln_index
• Lorenz curves: lorenz_curve
• Mean Log-Deviation: mean_logdev
• ModVR (Variation around the Mode): ModVR
• Pietra index (based on the Lorenz curve): pietra_lorenz
• Range (simple): range
• RanVR (Range around the Mode): RanVR
• Rao's quadratic entropy: rqe
• Rao's generalized quadratic entropy: qrqe
• Shannon heterogeneity: shannon
• Shannon entropy: shannon_entropy
• Simpson index (aka Herfindahl index): simpson
• Simpson dominance: simpson_dominance
• Theil index: theil
• Tsallis entropy: tsallis_entropy

Methods for Estimation

• Bootstrap sampling: div.bootstrap.bootci
  • State of the art approach is implemented: bias-corrected and accelerated [BCa] confidence intervals
  • Note this currently only works for measures of Type I (richness) heterogeneity
• Methods for which we have closed form estimators have those implemented by default

Measures to Add

• Accumulation curves: accumulation_curve

• Amato's index: amato

• Atkinson index: atkinson

• Average deviation: AvDev

• Birthday paradox estimate: bday_paradox_test

• Brillouin's D: brillouin_d

• Brillouin's E: brillouin_e

• Chao & Lee 1: chao_lee1

• Chao & Lee 2: chao_lee2

• Chao Presence/Absence: chao_pa

• Chao Quantitative: chao_quant

• Convex hull volume: hull_volume

• Dalton index: dalton

• Equal shares coefficient: equal_shares_coef

• Fisher's $\alpha$: fisher_alpha

• Functional Attribute Diversity: fad

• Functional Hill Numbers: func_hill

• Functional dendrogram: func_dendrogram

• Functional dispersion: func_disp

• Functional evenness: func_eve

• Functional logarithmic variance: func_logvar

• Functional unalikeability: func_unalikeability

• Functional Range: func_range

• Functional regularity: func_reg

• Functional richness: func_ric

• Functional variance: func_var

• Functional volume: func_vol

• Help index: help_index

• Huffman code length: huffman

• Leinster-Cobbold index: leinster_cobbold

• Lempel-Ziv code length: lempel_ziv

• Logarithmic variance: logvar

• Magurran-Henderson: magurran_henderson

• Margalef's index: margalef

• McIntosh's D: mcintosh_d

• McIntosh's E: mcintosh_e

• Michaelis-Menten: michaelis_menten

• Minimal majority: min_majority

• (MNDif) Mean absolute pairwise difference: MNDif

• Menhinick's index: menhinick

• Modified functional attribute diversity: mfad

• Pielou J: pielou

• Pooled Rarefaction: pooled_rarefaction

• Relative mean deviation: rel_mean_dev

• Numbers equivalent RQE: rqe_neq

• Rademacher complexity: rademacher

• Ricotta-Szeidl entropy: ricotta_szeidl

• Simpson's E: simpson_evenness

• Rarefaction (across samples): rarefaction

• Rarefaction (single sample): rarefaction_single

• Smith & Wilson B: smith_wilson_b

• Smith & Wilson 1D: smith_wilson_1d

• Smith & Wilson -lnD: smith_wilson_lnd

• StDev1 & StDev2: StDev1, StDev2

• Strong's index: strong_index

• Variance of logarithms: var_logs

• VarNC: VarNC

• We will also add wrappers for existing functions to account for synonymous indices such as

  • Hartley heterogeneity and "species richness": species_richness
  • and more