Notes

Outline

The Challenge of Missing and Uncertain Data in Information Visualization Cyntrica Eaton    Catherine Plaisant University of Maryland December 5, 2003

Sources of Missing and Uncertain Data Uncollected Data Refined Data Categories Data Source Confidentiality Non-Applicable Multivariate Combinations Statistical Sampling Small Samples Flawed Experimentation Estimated/Aggregated Data

HCIL Projects Visualizing Human-Related Statistical Data Categories are redefined over time Information not collected for certain groups in earlier periods Monitoring Oil Production Samples uncollected Oil wells added or closed Corrupted historical files Tracking Web Page Access Over a Period Pages added or deleted

Survey of Techniques

Taxonomy of Visualizations

Taxonomy of Visualizations

Taxonomy of Visualizations

Taxonomy of Visualizations

Taxonomy of Visualizations

Taxonomy of Visualizations

Taxonomy of Visualizations

Taxonomy of Visualizations

Slide 13

Slide 14

Slide 15

Slide 16

Slide 17

Slide 18

Slide 19

Slide 20

Slide 21

Slide 22

Slide 23

Slide 24

Slide 25

Slide 26

Slide 27

Slide 28

Issues DTD Define types of missing data information Link to explanations / more information Data File Generation Indicate missing data with applicable flags Visualization Tools Enforce constraints for user manipulations to preserve display integrity Dynamic Query Color indicators Provide rules for handling missing data values in calculations Minimize impact of large uncertainties

Data Representation

Data Representation

Data Representation

Issues DTD Define types of missing data information Link to explanations / more information Data File Generation Indicate missing data with applicable flags Visualization Tools Enforce constraints for user manipulations to preserve display integrity Dynamic Query Color indicators Provide rules for handling missing data values in calculations Minimize impact of large uncertainties

Impact of Uncertainty

Issues DTD Derive tags/attributes to encode missing data information Data File Generation Indicate missing data with applicable flags Tool Enforce constraints for user manipulations to preserve display integrity Dynamic Query Color indicators Provide rules for handling missing data values in calculations Minimize impact of large uncertainties

Future Work Complete implementation of missing/uncertain data indicators in TimeSearcher Conduct a Pilot Study to observe the effects of modifications on data interpretation Continue modifying TreeMaps to provide users with missing/uncertainty information

Future Work:  Pilot Study Observe the types of mistakes made in analyzing conventional displays Subject users to modified displays and observe mistakes Compare the difference between the two sessions