COMPEL Glossary / dimensionality-reduction
Dimensionality Reduction
Dimensionality reduction is a technique that simplifies complex datasets with many variables by identifying the most important underlying factors and representing the data in fewer dimensions.
What this means in practice
A dataset with hundreds of variables might be reduced to a handful of dimensions that capture the essential variation, making visualization, analysis, and downstream AI modeling more tractable. Common techniques include Principal Component Analysis (PCA) and t-SNE. For non-technical transformation leaders, dimensionality reduction is relevant because it helps AI systems focus on what matters in high-dimensional data and enables visualization of patterns that would otherwise be invisible. It is particularly useful in exploratory analysis during the Calibrate stage when organizations are assessing their data landscape.
Why it matters
Complex datasets with hundreds of variables can overwhelm both AI models and human analysts. Dimensionality reduction identifies the most important underlying factors, making visualization and analysis tractable. For business leaders, this translates to clearer insights from complex data, faster model development, and the ability to visualize patterns that would otherwise be invisible, supporting evidence-based decision-making.
How COMPEL uses it
Dimensionality reduction is particularly useful during the Calibrate stage when organizations are assessing their data landscape and need to understand patterns across many variables. It supports the cross-domain diagnostic that evaluates relationships between the 18 maturity domains. During Evaluate, dimensionality reduction can help identify the key factors driving or inhibiting transformation progress across multiple measurement dimensions.
Related Terms
Other glossary terms mentioned in this entry's definition and context.