The agriculture statistics platform.
AMMI biplots, GGE biplots, Diallel and Line by Tester, plus 20 experimental designs from CRD to Augmented to Alpha Lattice. With a plain-English AI explainer on the major outputs. Built for AICRP submissions and breeding decisions, not for thesis-writing alone.
| Source | df | SS | MS | F | p |
|---|---|---|---|---|---|
| Treatment | 4 | 42.18 | 10.55 | 38.4 | 0.0001 |
| Block | 2 | 3.22 | 1.61 | 5.87 | 0.0142 |
| Error | 8 | 2.20 | 0.27 | · | · |
| Total | 14 | 47.60 | · | · | · |
Treatment effect is highly significant (F = 38.4, p < 0.001). T3 (Hybrid) gave the highest grain yield at 6.82 t/ha, 62% above T2 (Old). Blocking absorbed real environmental variance. CV at 8.7% is within the typical range for field yield trials.
Four things you will not find anywhere else.
Other tools cover ANOVA. They do not cover the rest of the agronomist's and breeder's actual workflow. StatVeda starts where they stop.
AI explainer on the major outputs
The major designs (ANOVA, AMMI, GGE, Diallel, multiple regression, meta-analysis and more) ship an AI-written, plain-English interpretation. It reads only the visible numbers, never invents anything. Two sentences you can paste straight into a discussion section.
AMMI + GGE biplots, no code
Genotype × environment interaction analysis with proper biplots, point-and-click. Other browser tools do not ship biplots. R requires you to write the plotting code yourself.
20 designs, one consistent UI
CRD, RBD, Latin Square, Split-Plot, Split-Split, Strip-Plot, Factorial RBD, Augmented (CRD + RBD), Alpha + Balanced Lattice, AMMI, GGE, Diallel (Griffing 1-4), Line × Tester, Triple Test Cross. Same input format, same output style, same export buttons. Move between designs without re-learning the tool.
Beta is free, the rate locks later
Beta is the free 14-day trial. Paid plans open after beta, and the rate shown then is the rate beta users lock, it never goes up. Other tools nickel-and-dime per export or charge institutional rates upfront.
Built for agricultural research.
Plant breeders running multi-environment trials. Agronomists running split-plot field experiments. SAU and ICAR scientists writing AICRP results. Seed companies pooling pan-India trial data. One platform, one interface, every design.
Plant breeders
Agronomists
SAU + ICAR
StatVeda vs the tools you use today.
Honest comparison against the actual competitors.
| Feature | StatVeda | Free legacy tools | R / open-source | Paid desktop suites |
|---|---|---|---|---|
| Browser only, no install | ✓ | - | - | - |
| Mobile + tablet | ✓ | - | - | - |
| AMMI + GGE biplots, in browser | ✓ | - | code | - |
| Diallel (Griffing 1-4) | ✓ | - | code | - |
| Line × Tester, Triple Test Cross | ✓ | - | code | - |
| Augmented + Alpha + Balanced Lattice | ✓ | - | code | - |
| AI explainer on major outputs | ✓ | - | - | - |
| Auto interpretation paragraph | ✓ | - | - | - |
| Word + PPT + LaTeX export | ✓ | Word | - | Word/PDF |
| IJAS-compatible tables | ✓ | - | code | - |
| Free trial, no card | ✓ | Free | Free | - |
| Cost (single user, per year) | Free in beta | Free | Free, R req. | Rs 1L+ |
Comparison categories. Free legacy tools cover ANOVA + means well but lack browser delivery and biplots. R packages handle every method but require code, and biplots need extra plotting work. Paid desktop suites are full-featured but priced for institutions and locked to one machine.
From data to decision in three steps.
Pick your design
CRD, RBD, Latin Square, Split-Plot, Augmented, Alpha Lattice, AMMI, GGE, Diallel. Pick from the catalogue. The tool shows the data format it expects.
Enter your data
Grid entry (specify t and r), CSV upload, or paste a text block. Box-Cox, Yeo-Johnson, arcsine and logit transformations one click away.
Read your result
Instant ANOVA, treatment means with CLD letters, CD at 5% and 1%, biplots when applicable, AI explainer. Export to Word, PPT, LaTeX, PNG, SVG.
153 analyses across 9 categories, 137 verified.
Experimental designs, plant breeding, AMMI and GGE biplots, agronomy and pathology, classical tests, nonparametrics, regression, multivariate, and advanced models from mixed effects to survival. Every analysis is listed below, and so is what we are building next.
New (26): Scheffe Pairwise Comparisons, Contrasts and Polynomial Trends (one-way), Tukey HSD (all pairwise comparisons), McNemar Test (paired yes/no), Cochran's Q Test (k related yes/no treatments), Exact Binomial Test (one proportion), z Test for Proportions (one or two samples), Two Poisson Rates (rate ratio test), F Test for Two Variances / Chi-square Test of One Variance, G-test (likelihood ratio: goodness of fit or independence), Chi-square Goodness of Fit (expected ratio), Fleiss Kappa (agreement among many raters), Correlation CI and Comparison (Fisher z), Point-biserial Correlation, Runs Test for Randomness (Wald-Wolfowitz), Mood's Median Test (k groups), Kendall W (concordance among raters), Dunn Post Hoc Test (after Kruskal-Wallis), Conover-Iman Post Hoc Test (after Kruskal-Wallis), Nemenyi Post Hoc Test (after Friedman), Scheirer-Ray-Hare Test (two-way on ranks), Page L Trend Test (ordered treatments in blocks), Partial and Semi-partial Correlation, Kolmogorov-Smirnov Test (normality or two samples), Fligner-Killeen and Brown-Forsythe Tests (equal variances), Grubbs Test (single outlier)
Every analysis, by name
Open a group to see its full list. Each name opens that analysis.
Experimental Designs24
- Response Surface (CCD / Box-Behnken / quadratic fit)
- Dunnett (vs control) + Tukey-Kramer (unequal n)
- One Factor Analysis (CRD or RBD)
- CRD
- RBD
- Latin Square Design
- Two Factor Analysis (CRD / RBD / Split-Plot)
- Two Factor Factorial (RBD)
- Split-Plot Design
- Strip-Plot Design
- Three Factor Factorial (RBD or CRD)
- Augmented Design (RBD)
- Papadakis Nearest-Neighbour Adjustment
- Augmented Design (CRD, no blocks)
- Split-Split-Plot Design
- Balanced Lattice Design
- Alpha Lattice (PBIB) Design
- Pooled Analysis (one-factor over envs)
- Two-Factor over Environments (pooled)
- Multiple Comparisons (LSD, DMRT, Tukey)
- Multi-Character Batch ANOVA (one paste, many traits)
- Scheffe Pairwise Comparisonsnew
- Contrasts and Polynomial Trends (one-way)new
- Tukey HSD (all pairwise comparisons)new
Plant Breeding and Biplots25
- AMMI (Additive Main + Multiplicative Interaction)
- GGE Biplot (Genotype + GxE)
- Path Analysis from a pre-computed correlation matrix
- Path Analysis on Genotypic / Phenotypic / Environmental correlations
- Stability Analysis (Eberhart-Russell)
- Generation Means (additive-dominance, simple)
- Generation Means (3/5/6-parameter + Cavalli JST)
- Line x Tester Analysis
- Diallel Analysis (Griffing all 4 methods)
- Partial Diallel Analysis
- Triple Test Cross Analysis
- Probit Analysis (LC50 / LD50 / GR50)
- Drought-Tolerance Indices (SSI / TOL / MP / STI / GMP / YI / YSI / DRI)
- Genetic Variability (GCV / PCV / heritability / genetic advance)
- D² Genetic Divergence (Mahalanobis + Tocher clustering)
- Selection Index (Smith-Hazel discriminant function)
- Stability: Wricke Ecovalence + Shukla Variance
- Heterosis + Inbreeding Depression
- Yield-Component Analysis (correlation + path coefficients)
- Weed-Control Indices (WCE / WI / CRE / WPE)
- Horsfall-Barratt Disease Severity + AUDPC
- Field Layout Generator (CRD / RBD / Latin / Augmented / Alpha lattice)
- GGE Which-Won-Where Polygon (Yan-Hunt 2000)
- GGE Mean vs Stability (Yan-Kang 2003)
- GGE Ideal-Genotype Ranking (Yan-Tinker 2006)
Classical Tests26
- Comparing Means (one-sample / two-sample / paired / one-way ANOVA)
- One-Sample t-test
- Two-Sample t-test
- Paired t-test
- Chi-square Test
- Mann-Whitney U Test
- Welch ANOVA + Games-Howell
- Effect Size Calculator (d / g / eta2 / omega2 / V / OR / RR)
- Sample Size + Power (t / ANOVA / chi-square / regression)
- Fisher Exact + Freeman-Halton (small contingency tables)
- Cochran-Mantel-Haenszel (stratified 2x2xK)
- TOST Equivalence Test (two one-sided)
- Weighted Cohen Kappa (rater agreement)
- Intraclass Correlation (ICC, all 6 forms)
- Lin's Concordance (CCC) + Bland-Altman
- McNemar Test (paired yes/no)new
- Cochran's Q Test (k related yes/no treatments)new
- Exact Binomial Test (one proportion)new
- z Test for Proportions (one or two samples)new
- Two Poisson Rates (rate ratio test)new
- F Test for Two Variances / Chi-square Test of One Variancenew
- G-test (likelihood ratio: goodness of fit or independence)new
- Chi-square Goodness of Fit (expected ratio)new
- Fleiss Kappa (agreement among many raters)new
- Correlation CI and Comparison (Fisher z)new
- Point-biserial Correlationnew
Nonparametric16
- Kruskal-Wallis (k groups)
- Jonckheere-Terpstra (ordered trend, k groups)
- Durbin Test (incomplete block design)
- Wilcoxon Signed-Rank (paired)
- Friedman (k matched samples)
- Spearman Rank Correlation
- Kendall tau-b
- Sign Test (paired)
- Runs Test for Randomness (Wald-Wolfowitz)new
- Mood's Median Test (k groups)new
- Kendall W (concordance among raters)new
- Dunn Post Hoc Test (after Kruskal-Wallis)new
- Conover-Iman Post Hoc Test (after Kruskal-Wallis)new
- Nemenyi Post Hoc Test (after Friedman)new
- Scheirer-Ray-Hare Test (two-way on ranks)new
- Page L Trend Test (ordered treatments in blocks)new
Regression and Correlation12
- Simple Linear Regression
- Nonlinear Growth Curve (Logistic / Gompertz / von Bertalanffy / Richards)
- Quantile Regression
- Penalized Regression (Ridge + LASSO with cross-validation)
- Stepwise Regression
- Path Analysis (simple, standardised regression)
- Multiple Regression
- Mediation Analysis (Hayes / Sobel + bootstrap)
- Zero-Inflated Poisson + Hurdle (count data with excess zeros)
- Correlation (Pearson + Spearman + Kendall)
- Pearson Correlation Matrix
- Partial and Semi-partial Correlationnew
Multivariate8
- Exploratory Factor Analysis (PAF + varimax + Horn parallel)
- Principal Component Analysis
- K-Means Cluster Analysis
- Hierarchical Clustering
- Principal Coordinate Analysis (classical MDS)
- Cluster Validation (silhouette / Calinski-Harabasz / Davies-Bouldin)
- Correspondence Analysis (contingency table)
- Canonical Correlation Analysis
Advanced Models30
- Meta-Analysis (fixed and random effects, forest plot)
- Cronbach's α (reliability)
- Power & Sample Size
- ANCOVA (one covariate)
- Logistic Regression (binary)
- Repeated Measures ANOVA
- Control Charts (X̄/R, X̄/S, I-MR, p, c, u)
- Kaplan-Meier Survival + Log-Rank
- MANOVA (multivariate ANOVA)
- Time Series ACF / PACF
- Mann-Kendall Trend Test + Sen's Slope
- Compound Growth Rate (log-linear)
- Crop Growth Analysis (CGR / RGR / NAR / LAI / LAD)
- ADF Stationarity Test (Augmented Dickey-Fuller)
- KPSS Stationarity Test (level or trend)
- Granger Causality (X causes Y, F-test)
- SARIMA(p,d,q)(P,D,Q)_s, seasonal ARIMA
- Poisson Regression (counts)
- Negative Binomial Regression (overdispersed counts)
- Post-hoc P-value Corrections (Bonferroni / Holm / BH FDR)
- Linear Discriminant Analysis (Fisher / Canonical)
- Multinomial Logistic Regression (3+ categories)
- PLS-SEM (Partial Least Squares Structural Equation Modelling)
- Confirmatory Factor Analysis (SEM subset)
- ARIMA(p, d, q), Box-Jenkins
- Cox Proportional Hazards (survival regression)
- Generalised Linear Mixed Model (Gaussian / Binomial / Poisson)
- Three-Level Hierarchical Linear Model (REML)
- Linear Mixed-Effects (random intercept)
- Ordinal Logistic Regression (proportional odds)
Normality and Diagnostics9
- Variance Homogeneity (Levene + Bartlett)
- Data Transformations (Box-Cox / Yeo-Johnson / Arcsine / Logit)
- Shapiro-Wilk Normality
- Anderson-Darling Normality
- Jarque-Bera Normality
- Q-Q Plot (normal)
- Kolmogorov-Smirnov Test (normality or two samples)new
- Fligner-Killeen and Brown-Forsythe Tests (equal variances)new
- Grubbs Test (single outlier)new
In development (75)
These are being built now. An analysis is released only after its results match an independent Python reference, so this list shrinks as the one above grows.
Nested (hierarchical) ANOVA, RBD with missing plots, Tukey test for non-additivity, 2 x 2 crossover design, Graeco-Latin square, Balanced incomplete block design, 2^k factorial with Yates effects, Relative efficiency of designs
Polynomial regression, Weighted least squares, Robust (Huber) regression, Nonlinear regression (exponential, power, Michaelis-Menten, Mitscherlich), Linear and quadratic plateau models, Log-logistic dose response (ED50), Quadratic response and economic optimum dose, Cobb-Douglas production function, Chow structural break test, Theil-Sen slope, Deming regression, LOWESS smoothing, Binomial GLM (logit, probit, cloglog), Gamma GLM, Partial least squares regression
Hotelling T-squared, Box M test, Mahalanobis outlier detection, Quadratic discriminant analysis, KMO and Bartlett sphericity, Henze-Zirkler multivariate normality, Distance and similarity matrices, Cophenetic correlation, DBSCAN clustering
Hardy-Weinberg equilibrium test, Parent-offspring heritability, Genotypic and phenotypic correlations, North Carolina Design I, North Carolina Design II, Finlay-Wilkinson regression, AMMI stability value, Lin and Binns superiority measure, Francis-Kannenberg CV grouping, Nassar-Huehn nonparametric stability, Kang rank-sum, Response to selection, PIC and marker diversity, Nei genetic distance, Two-point linkage analysis, Degree of dominance and inbreeding depression, Genotype BLUPs and heritability, Multi-environment variance components
Land equivalent ratio and intercropping indices, Intercrop competition indices, Nutrient use efficiency, Growing degree days, Seed germination indices, Irrigation water quality (SAR, RSC), AUDPC and infection rate, Abbott and Henderson-Tilton corrected efficacy, Spatial dispersion indices, Species diversity indices, Site similarity indices
Benefit-cost analysis of treatments, Partial budget and marginal rate of return, NPV and IRR, Instability indices (Cuddy-Della Valle), Survey sample size, Stratified sampling, Exponential smoothing, Seasonal decomposition, Ljung-Box test, Pettitt change-point test, Distribution calculator, Distribution fitting, ROC curve and AUC, Diagnostic accuracy
Where every number comes from.
No approximations, no hardcoded guesses. Every critical value derives from textbook sources and is cross-checked against canonical reference implementations to machine precision.
F-distribution p-values
Regularized incomplete beta with the symmetry transform; within 4e-10 of SciPy on a test grid
t critical values (CD)
Solved exactly from the t distribution for any df and α (within 3e-11 of SciPy); a few older tools still read a 3-decimal table
Studentized Range q (Tukey, DMRT)
Numerical integration for any α and df in Tukey-Kramer and Tukey HSD (within 1e-8 of SciPy); the Multiple Comparisons tool uses tabulated values
LSD / CD(5%) / CD(1%)
t(df, α) × √(2 MSE / r), exact formula, no approximation
DMRT critical ranges
R_p = q(α, p, df) × √(MSE/r), Studentized Range table, p = 2..k
Tukey HSD
HSD = q(α, k, df) × √(MSE/n_h), proper Studentized Range critical value
Eberhart-Russell + PJ stability
bi and s²d per E-R 1966; PJ index per Perkins-Jinks 1968
Griffing diallel (Methods 1-4, Models I+II)
All four Griffing 1956 methods, fixed and random; GCA, SCA, reciprocal effects
AMMI biplot (IPCA1 vs IPCA2)
Gauch σ^λ partitioning, λ_g=1, λ_e=0; cross-checked to canonical reference implementation
GGE biplot (PC1 vs PC2)
Yan and Kang 2003 environment-centred SVD, λ=0.5 symmetric per Yan and Hunt 2000
Augmented designs (CRD + RBD)
Federer 1956; CD(check vs check), CD(check vs test), CD(test vs test) all distinct
Mixed-effects + REML
Profile-likelihood REML for variance components, matched to statsmodels; between-within df for fixed effects
Start the 14-day trial.
14-day trial, no card, live now. Paid plans open after beta: Rs 499 per month for individuals, Rs 2,499 per month for labs, from Rs 50,000 per year for institutions. Sign up during beta and you lock that rate when they launch.
SAU + ICAR DISCOUNT ON REQUEST · IJAS-COMPATIBLE OUTPUT · REST API ON INSTITUTION PLAN