Move from dashboard charts to defensible statistical analysis
This page previews the statistical analysis workspace we're building: weighted estimates, significance, regression, experiments, reliability, factor structure, key drivers, and trends, with assumptions and calculation details exposed rather than hidden.
Recommended test
Weighted proportion comparison
Comparing satisfaction between two regions with survey weights and unequal sample sizes - illustrative output, not a computed result.
None of the analyses below are implemented yet. The cards describe the statistical engine on our roadmap; only the driver demo further down is a working interactive preview.
Significance & intervals
Proportions, means, t-tests, chi-square, ANOVA, confidence intervals, and effect sizes.
Regression
Linear, logistic, and ordinal models with readable diagnostics and exportable specifications.
Survey design
Weights, strata, clusters, PSU, design effects, effective sample size, and weighted estimates.
Reliability & factors
Cronbach alpha, item-total diagnostics, exploratory factor analysis, and scale review.
Key drivers
Estimate which experience attributes are most strongly associated with NPS, CSAT, retention, or other outcomes.
Experimental analysis
Treatment-control comparisons, uplift, confidence intervals, assignment metadata, and predefined outcomes.
Identify the strongest improvement opportunity
Move a slider - the result recalculates live. This demo only picks the lowest of the three scores; it does not run a real key-driver regression.
Show the math and the data lineage
- View assumptions and model settings
- See included/excluded cases
- Trace weights and filters
- Export analysis-ready scripts
- Link narrative conclusions to supporting outputs
Roadmap items - none of this is available in the product today.