FactorLens

About

Turn a small numeric dataset into an explanation you can inspect and a forecast you can understand. FactorLens is built for business users, students, and researchers who want rigorous multifactor analysis without sending data to a server. Start with the guided Beginner Mode, then type observations, paste a spreadsheet table, or import CSV/TSV. Analyze up to 10 numeric factors and 1,000 rows directly on your iPhone or iPad. EXPLAIN THE RESULT • Multiple linear regression estimates each factor's association with the outcome while holding the other factors constant. • Standardized effects make factors easier to compare. • Exact Shapley/LMG importance fairly distributes the model's explained variance across correlated factors. • Confidence intervals and p-values keep uncertainty visible. CHECK THE MODEL • Residual and Q-Q charts reveal patterns and unusual distributions. • VIF and factor-correlation diagnostics highlight multicollinearity. • Leverage, studentized residuals, and Cook's distance identify influential observations. • A Breusch-Pagan check flags changing residual variance. MAKE A TRANSPARENT FORECAST FactorLens compares ordinary least squares with standardized Ridge regression using deterministic nested cross-validation. A simpler model is preferred when its predictive error is statistically indistinguishable. Predictions show the selected method, uncertainty or validation error, and an extrapolation warning when inputs fall outside the observed range. PRIVATE BY DESIGN Studies and calculations stay on your device. FactorLens has no account, advertising, analytics, tracking, or cloud service. Export the original data, fitted values, and factor results as CSV, or share a concise PDF report. FactorLens describes statistical associations, not causal effects. Results depend on data quality, model assumptions, and the variables you choose.