One row per customer is a different problem from one row per customer per month. Shape decides which methods can apply to your file at all. More are in build.
That is the actual report this analysis produces — not a mock-up. Every analysis on this page delivers the same shape of output on your own data.
Every analysis classified under this data shape.
Should I accept this lot?
Do two measurement methods agree well enough to be used interchangeably?
Are these categories related?
Library tool, v2: logistic regression of a two-class outcome on one to eight drivers with a stratified held-out split; AUC with a 95% i
What natural segments exist in my data?
Library tool, v2: pairwise Pearson (or Spearman) correlation across two to eight numeric measures with a 95% interval and p-value per p
How do two categorical variables map together? Biplot of category associations — brands vs attributes, segments vs behaviors.
What is associated with my event counts?
What's in my dataset?
Is my data normal, and are there outliers?
What distribution does my data follow? Fit and compare candidate distributions with goodness-of-fit tests.
How does response change with dose or intensity? Nonlinear curve fitting with EC50/IC50 estimates.
What latent factors underlie my metrics?
Which predictors does the model actually depend on?
What matters most that we do worst? Importance vs performance quadrants from survey data for prioritization.
Which regression family fits my outcome — and what drives it?
Where does my data's family tree naturally split?
What is the causal effect when treatment is confounded? Two-stage least squares with instrument diagnostics.
Categorical inter-rater agreement; complements ICC.
What moves satisfaction most?
If I only contact the top N%, how much of the value do I capture?
Ordinal survey items done right.
Library tool, v2: ordinary least squares of one numeric outcome on one to five drivers; eight objects into run_objects (verdict, headli
Does X drive Y through M; conditional effects.
Pool effect sizes across studies.
How bad is my missing data and does it bias my results? Missingness patterns, MCAR checks, imputation impact.
What is our Net Promoter Score, and is it moving or differing by segment?
Which few causes drive most of the problem?
How many real dimensions does my data have?
Is the process capable of meeting spec limits.
Did crossing the threshold cause the change?
Is my survey scale reliable?
Where should I set the cutoff given MY costs?
What hidden segments exist in categorical responses? Model-based clustering for survey and choice data.
Fit several classifiers/regressors on one dataset, compare honestly (CV, metrics, stability).
What drives a 1-to-5 rating outcome? Proportional-odds models for ordered outcomes instead of pretending they are continuous.
What drives the WORST and BEST cases, not the average? Regression at chosen quantiles of the outcome.
Does my sample look like my population? Compute and apply post-stratification weights, compare weighted vs raw estimates.
Heterogeneous treatment effects; persuadables vs sure-things.
Tell us what went wrong, in your own words. We capture the page you're on automatically, so no need to describe where you are.