ggEDA: Visualisations for exploratory data analysis using tiled one-dimensional graphics and parallel coordinate...

Area of Science:

• Data Science
• Bioinformatics
• Computational Biology

Background:

• Exploratory Data Analysis (EDA) is crucial for identifying data quality issues and generating hypotheses.
• Analyzing multidimensional relationships in datasets often requires significant coding and statistical expertise.
• Current methods for exploring complex data patterns can be time-consuming and require manual inspection.

Purpose of the Study:

• To introduce the ggEDA R package for streamlining multidimensional data exploration.
• To provide accessible visualization tools for uncovering multi-feature relationships and data quality issues.
• To develop an interactive web application (interactiveEDA) for non-programmers to perform data exploration.

Main Methods:

• Development of the ggEDA R package featuring interactive parallel coordinate plots (PCPs) for large, quantitative datasets.
• Introduction of tiled one-dimensional plots in ggEDA for revealing missingness and categorical relationships in smaller datasets.
• Creation of the interactiveEDA web app to enable non-programmers to explore data patterns.

Main Results:

• ggEDA effectively reduces the coding effort and time needed to detect multi-feature relationships.
• PCPs in ggEDA are suitable for large datasets with quantitative features.
• Tiled plots in ggEDA excel at visualizing missing data and categorical associations in smaller datasets.
• interactiveEDA provides a user-friendly interface for interactive data pattern interpretation.

Conclusions:

• The ggEDA package and interactiveEDA web app significantly simplify and accelerate multidimensional data exploration.
• These tools democratize data analysis by reducing the need for extensive programming or statistical modeling.
• ggEDA and interactiveEDA facilitate the discovery of data quality issues and underlying patterns, aiding hypothesis generation.