Dr. Semiha B. Sekerli, Ph.D.QUANTITATIVE METHODOLOGIST
Welcome to🐈⬛ Pearl
A Web-Based Statistics Tool for Doctoral Researchers
Pearl is Chico's little brother — Dr. Semiha B. Sekerli's open-access quantitative methodology tool. Upload your data, pick a test, and Pearl will run the analysis and interpret the results in plain language. Built for doctoral students and researchers who want SPSS-level analysis without the price tag — and your data never leaves your browser.
CSV files only for now. Pearl will detect each column's type and show you a preview before you pick a test.
Your data stays in your browser. Nothing gets uploaded to a server. All calculations happen on your device.
Data Preview
Pick an analysis
Advanced · Python in your browserFirst click downloads ~30 MB (numpy, scipy, statsmodels). Runs entirely on your machine — data never leaves the browser.
No options needed
Pearl will summarize every column. Quantitative columns get mean/SD/median/min/max. Categorical columns get frequency counts.
Configure t-test
Configure correlation
Configure One-Way ANOVA
Configure Chi-Square Test of Independence
Configure Simple Linear Regression
Configure Multiple Regression
Configure Logistic Regression
Categorical predictors are dummy-coded automatically (first level = reference).
Configure MANOVA
Configure Exploratory Factor Analysis
Configure ANCOVA
Configure Factorial (Two-Way) ANOVA
Configure Paired t-Test
Each row should be one subject — Pearl pairs Variable 1 and Variable 2 by row.
Configure Mann-Whitney U Test
Configure Wilcoxon Signed-Rank Test
Each row should be one subject — Pearl pairs the two variables by row. Non-parametric alternative to the paired t-test.
Configure Kruskal-Wallis Test
Pearl runs Dunn's post-hoc with Bonferroni correction automatically when the omnibus test is significant.
Configure Cronbach's α
Items should be scored in the same direction. If an item is reverse-coded, recode it before uploading.
Configure Repeated-Measures ANOVA
Each row should be one subject. Pearl treats your selected columns as repeated measurements on the same person (e.g., pretest, midtest, posttest).
Configure Multinomial Logistic Regression
Categorical predictors are dummy-coded automatically. Each non-reference outcome level gets its own set of coefficients.
🐈⬛ Pearl is loading Python...
First-time setup downloads the scientific Python stack. After this, advanced tests run instantly.
Pyodide runtime (~6 MB)
numpy · scipy · pandas (~20 MB)
statsmodels (~5 MB)
Warming up the interpreter
Your data never leaves your browser. All computation happens locally.
💤
🤖 Knock knock! Hi! I'm Pearl — Dr. Sekerli's stats tutor. Need help picking a test or interpreting results? Click me!