Paste your field or laboratory data, choose the design you actually used, and get a publication-ready analysis — the ANOVA, the means separation, the charts and a written interpretation. Seventy-three modules covering experimental designs, biometrical genetics, multivariate methods, non-parametric tests, econometrics and time series.
Every module takes data in the layout you already keep it in, and returns the tables a journal expects — not a console dump you then have to format.
CRD, RBD and Latin square; one, two and three factor analyses; split and strip plot; balanced and alpha lattice; augmented and group balanced block designs; repeated measures; pooled analysis over years or locations; analysis of covariance; and randomised field layouts.
Genetic variability with GCV, PCV, heritability and genetic advance; Mahalanobis D² divergence with Tocher clustering and dendrograms; heterosis; North Carolina designs I, II and III; generation means; line × tester; Griffing, Hayman and Gardner–Eberhart diallels; triple test cross; path analysis; stability; selection indices.
Principal component analysis, factor analysis, discriminant analysis, MANOVA, cluster analysis and k-means, probit and logit, count data regression, and response surface methodology.
Frequency distributions, cross-tabulation with association measures, means comparison and t-tests, correlation, and linear and multiple regression.
Chi-square, Mann–Whitney, Wilcoxon, Kruskal–Wallis, Friedman, Mood's median, sign test, runs test, Spearman's rank correlation and Kolmogorov–Smirnov.
Cobb–Douglas and CES production functions, stochastic frontier analysis, growth and instability, crop decomposition, Hazell's decomposition, the policy analysis matrix, financial feasibility, linear programming, and adoption logit/probit and tobit models.
Moving averages and rolling compound growth rates, ARIMA and SARIMA forecasting, Holt–Winters smoothing, seasonal decomposition, spatial market integration, Granger causality, vector autoregression, ARCH/GARCH volatility and unit root testing.
Each module carries its own theory page: the model, the estimators, the degrees of freedom, the standard errors, how to read the output, and the worked example it was validated against.
No installation, no licence key, no syntax to learn. Four steps from spreadsheet to manuscript table.
Copy straight out of Excel. Every module shows a worked sample dataset in the exact layout it expects, so you can check the shape before you commit.
Treatments, replications, locations, characters. Pick the post-hoc test and transformation you need. Nothing is guessed on your behalf.
Compact publication-style tables, significance marks, critical differences, interactive charts, and a written interpretation of what the numbers actually say.
Excel, Word, PDF or plain text, formatted and ready to drop into a thesis chapter or a journal submission.
A statistics package is only worth as much as its arithmetic. Every module in AGSTAT has been driven end to end and its output compared, value by value, against established reference implementations.
Over 620 individual values — sums of squares, mean squares, F ratios, effects, standard errors, cluster memberships — reconciled with the worked examples in the standard biometrical texts and with independent computations carried out to full double precision.
Critical values come from exact t, F, chi-square and studentised range quantiles computed to about fifteen significant figures, not from rounded printed tables.
Where two defensible conventions exist, the one used is named on the theory page with its citation, so a reviewer can check what was done.
Datasets are analysed in the request and are not retained after the result is rendered. Accounts are protected by session authentication and a hardened server configuration.
Start on the free plan and pay only when you need the full library. Payment is by UPI or bank transfer — upload the receipt and an administrator activates your account.
One analysis is counted each time you submit a dataset. Re-reading a result you have already produced does not count again. Unused analyses carry over when you renew, and renewing early never costs you days.
Built by the AgriSaarthi Portal for the analyses agricultural research actually runs.
Analyse thesis trials without buying a licence or learning a syntax. The theory page beside each module explains what the test is doing and why.
Combining ability, heterosis, divergence, stability and selection indices in one place, on the mating designs breeding programmes actually use.
Production functions, decomposition, market integration and forecasting, with the diagnostics a referee will ask for.
Multi-location and multi-year pooled analyses, covariance analysis and repeated measures, with output that goes straight into a report.
One-factor analysis with the full ANOVA and treatment means, at no cost and with no time limit. Upgrade when your work needs the rest of the library.
Sign in or create an account