ANALYSIS TOOL BOX

8 functions

Descriptive Analytics

Clustering, dimensionality reduction, and pattern discovery.

Understand the structure of your data — who groups together, what drives variance, and how to create balanced comparison groups.

Highlight functions

CreateKMeansClusters

K-means clustering with automatic visualization of cluster assignments. Returns the labeled DataFrame and a cluster profile summary.

from analysistoolbox.descriptive_analytics import CreateKMeansClusters

df_clustered = CreateKMeansClusters(
    dataframe=df,
    list_of_clustering_variables=["recency", "frequency", "monetary"],
    number_of_clusters=4
)

CreateHierarchicalClusters

Agglomerative hierarchical clustering with dendrogram output. Useful when you don't know the number of clusters in advance.

from analysistoolbox.descriptive_analytics import CreateHierarchicalClusters

CreateHierarchicalClusters(
    dataframe=df,
    list_of_clustering_variables=["feature_1", "feature_2", "feature_3"]
)

ConductPropensityScoreMatching

Creates matched control and treatment groups using propensity scores — the standard approach for causal inference from observational data.

from analysistoolbox.descriptive_analytics import ConductPropensityScoreMatching

df_matched = ConductPropensityScoreMatching(
    dataframe=df,
    treatment_variable="received_treatment",
    list_of_covariates=["age", "income", "baseline_score"]
)

ConductPrincipalComponentAnalysis

PCA with scree plot, explained variance, and a 2D projection of the data.

All functions

| Function | Description | |---|---| | ConductManifoldLearning | Non-linear dimensionality reduction (UMAP, t-SNE) | | ConductPrincipalComponentAnalysis | PCA with visualization | | ConductPropensityScoreMatching | Balanced groups for causal inference | | CreateAssociationRules | Market basket / affinity analysis | | CreateGaussianMixtureClusters | Soft probabilistic clustering | | CreateHierarchicalClusters | Hierarchical clustering with dendrogram | | CreateKMeansClusters | K-means clustering | | GenerateEDAWithLIDA | AI-powered EDA using Microsoft LIDA |