Color Regression Lines
in Data Desk

What is DataDesk?

DataDesk, developed by Paul Velleman in the 1980s, was a pioneering tool for exploratory data analysis and visualization. Its groundbreaking feature was the use of dynamic, linked graphics that enabled users to interactively explore their data. This innovation set DataDesk apart from contemporary software packages like SPSS. To grasp the concept of dynamic graphics, it's helpful to contrast it with static graphics:

Static Graphics

• Graphs are not linked to other objects (e.g., data tables)
• Output is fixed: what you see is what you get; you can't click on data points for more information

Dynamic Graphics

• Graphical output is linked to all other objects
• Allows for real-time manipulation and exploration: Facilitates rapid exploration of different data views, pattern identification, and hypothesis generation

Example: Color Regression in Dynamic Visualization

Color regression illustrates the power of dynamic data visualization:

• Model: Four variables - posttest scores (dependent), pretest scores, cognitive test scores, and integration test scores (independent)
• Interactive color assignment: Observations are grouped into four colors based on clustering patterns
• Result: Four sub-models (represented by different colored regression lines)

Key finding: The relationship between pretest and posttest scores varies based on cognitive ability:

• Low cognitive ability: Negative relationship
• Medium or high cognitive ability: Positive relationship

The Legacy of DataDesk

While DataDesk was discontinued on June 30, 2024, its influence persists. Modern tools like JMP, SAS Viya, and Tableau carry forward the tradition of interactive, dynamic data visualization.

Conclusion

Data visualization remains a crucial component of data science. The principles pioneered by DataDesk continue to shape the field, emphasizing the importance of interactive, exploratory approaches to data analysis.

Last updated: August 5, 2024