Context before conclusions
Descriptive summaries
Compute counts, missing values, mean, standard deviation, range, and quartiles. Non-numeric fields and incomplete observations remain visible rather than silently disappearing.
Correlation inspection
Review pairwise numerical associations. Correlation describes a relationship in the selected observations; causal claims require an appropriate study design and assumptions.
Data quality
Preview imported rows, inferred types, and missingness. Preserve original CSV content and its hash so the dataset underlying an analysis remains identifiable.
Diagnostics and limitations
Inspect residuals and held-out predictions after regression. Small, constant, or degenerate datasets produce useful limitations instead of an invented performance claim.
The initial release focuses on descriptive statistics, correlation, regression, and diagnostics. Formal confirmatory testing and complex repeated-measures analyses require an appropriately specified method and are not presented as one-click conclusions.