Why Gyanank exists
The problem
India produces an enormous amount of research every year — Pharm.D, MD, MS, MDS and nursing theses, PhD work, faculty papers, NGO surveys, field studies in tier-2 and tier-3 cities. Almost all of it starts the same way: a Google Form, an Excel sheet, and a person who has never been taught statistics.
Then comes the wall. They can't run the test. SPSS assumes you're already an expert and charges accordingly. Jamovi and JASP are free, but a first-time researcher still can't pick the right test or read the output. ChatGPT and other AI chatbots will talk about statistics, but uploading patient or survey data to them is a data breach waiting to happen — and most guides and institutions won't accept AI-generated analysis anyway. None of these tools start where the researcher starts: at data collection.
So people do what they can. They pay a statistician ₹20,000 or more — money a student doesn't have — and wait. They get back tables they can't explain in a viva. Some publish conclusions they don't understand. Some get their thesis sent back and lose months. Some decide research isn't for them and never try again.
The talent is everywhere in India. The tools aren't. That's the gap Gyanank is built to close.
Our mission
Gyanank exists so that no student, clinician, teacher or field worker in India has to pay ₹20,000 and wait weeks just to understand their own data.
We are building one simple place where you design your form, collect responses, run the right statistical test and get a report you can actually defend — with every step explained in plain words. Indian pricing. Indian languages. Your data protected under DPDPA.
Two words sit at the heart of everything: access and simplicity.
Checked against R
Gyanank's statistics run in Python, on SciPy and statsmodels, open libraries used in published research around the world. To show that its answers are the ones a statistician would get, we analyse the same data in Gyanank and in R, the statistics software many universities and journals use, and compare every figure.
Last checked on 04-10-2026 with R 4.6.1: 22 of 22 results are the same.
See every result
| Test | Gyanank | R |
|---|---|---|
| Student's t-test | t(78) = -2.01, p = .048, d = 0.448 | t(78) = -2.01, p = .048, d = 0.448 |
| Welch's t-test | t(45.0) = -3.19, p = .003, d = 0.713 | t(45.0) = -3.19, p = .003, d = 0.713 |
| Paired t-test | t(79) = 2.55, p = .013, d = 0.285 | t(79) = 2.55, p = .013, d = 0.285 |
| One-way ANOVA | F(2, 77) = 7.84, p = .001, η² = 0.169 | F(2, 77) = 7.84, p = .001, η² = 0.169 |
| Mann–Whitney U | U = 591.0, p = .043, r = 0.261 | U = 591.0, p = .043, r = 0.261 |
| Kruskal–Wallis | H(2) = 11.52, p = .003, ε² = 0.124 | H(2) = 11.52, p = .003, ε² = 0.124 |
| Wilcoxon signed-rank | W = 736.5, p = .011, r = 0.284 | W = 736.5, p = .011, r = 0.284 |
| Chi-square, 3 × 3 | χ²(4) = 6.51, p = .164, V = 0.202 | χ²(4) = 6.51, p = .164, V = 0.202 |
| Chi-square, 2 × 2 (Yates) | χ²(1) = 0.11, p = .740, V = 0.037 | χ²(1) = 0.11, p = .740, V = 0.037 |
| Fisher's exact test | p = .432, OR = 2.71 | p = .432, OR = 2.71 |
| Pearson correlation | r = -0.258, p = .021 | r = -0.258, p = .021 |
| Spearman correlation | ρ = 0.245, p = .029 | ρ = 0.245, p = .029 |
| McNemar, chi-square form | p = .001, OR of change = 4.333 | p = .001, OR of change = 4.333 |
| McNemar, exact form | p = .146, OR of change = 3.000 | p = .146, OR of change = 3.000 |
| Linear regression, model | F(2, 77) = 10.86, p < .001, R² = 0.220, adjusted R² = 0.200 | F(2, 77) = 10.86, p < .001, R² = 0.220, adjusted R² = 0.200 |
| Linear regression, hours | β = 0.980, SE = 0.214, p < .001 | β = 0.980, SE = 0.214, p < .001 |
| Linear regression, sleep | β = 0.137, SE = 0.385, p = .723 | β = 0.137, SE = 0.385, p = .723 |
| Logistic regression, hours | OR = 1.80 [1.19, 2.70], p = .005 | OR = 1.80 [1.19, 2.70], p = .005 |
| Logistic regression, sleep | OR = 1.90 [0.93, 3.90], p = .079 | OR = 1.90 [0.93, 3.90], p = .079 |
| Cronbach's alpha | α = 0.835 | α = 0.835 |
| Shapiro–Wilk normality check | p = .382 | p = .382 |
| Levene equal-variance check | p = .164 | p = .164 |
The test data is a made-up study of 80 respondents, read by both. Where R's default settings differ from Gyanank's by definition, such as which odds ratio Fisher's exact test reports, R was set to Gyanank's method.