Correlation Analysis

Correlation analysis is a key feature of every Visplore version, including Visplore Free.

Correlations between variables play a key role in data science, engineering, and many other fields. The purpose of correlation analysis is manifold and ranges from data exploration and feature selection to understanding sensitivities and doing root-cause analysis of industrial process data.

Challenge

For real data, however, correlation analysis is typically more challenging than computing some well-known statistics. Outliers may introduce bias. Even worse, if the data consists of multiple clusters, disregarding them will likely lead to misleading results. For example, changes to an industrial process may affect how sensors correlate with each other and require that correlations are computed separately before and after the change. In general, data preparation and result validation are important when analyzing correlations and may account for the majority of the overall effort.

Solution

All Visplore versions feature powerful interactive tools for correlation analysis. These tools will help you to discover and model correlations while minimizing the effort for data preparation and result validation.

Correlation analysis in Visplore

Highlights

  • Discover correlations among dozens of variables within seconds
  • Understand correlations with a specific target variable
  • Access correlation coefficients and p-Values any time
  • Validate correlations in scatterplots and time line views with one click
  • Refine your results by selecting coherent, plausible data – all correlations update immediately
  • Break down complex correlations by unlimited drill-down possibilities
  • Fit and export regression functions