DataMeadow A Visual Canvas for Analysis of Large-Scale Multivariate Data Niklas Elmqvist  (elm@lri.fr) –  John Stasko  (st...
In media res… <ul><li>Let’s start with a demo! </li></ul>
Parallel Coordinates <ul><li>First proposed by Alfred Inselberg in  The Visual Computer  in 1985 </li></ul><ul><li>Basic i...
Elements and Dependencies <ul><li>Visual Elements </li></ul><ul><ul><li>Conforms to system-wide data format </li></ul></ul...
The DataRose <ul><li>2D starplot display </li></ul><ul><li>Transformer  visual element </li></ul><ul><li>Average as black ...
DataRose Representations <ul><li>Parallel coordinate mode </li></ul><ul><ul><li>See all details </li></ul></ul><ul><ul><li...
DataRose Types <ul><li>DataRoses can be of several different types </li></ul><ul><ul><li>Each type represents a specific m...
Viewers and Annotations <ul><li>Viewers are sink elements: accept input - no output </li></ul><ul><li>Shows quantitative i...
Evaluation <ul><li>Evaluation : expert review (think-aloud protocol) </li></ul><ul><li>Participants : two visualization re...
Contributions <ul><li>A highly interactive visual canvas ( DataMeadow ) for multivariate data analysis using multiple smal...
Future Work <ul><li>Additional visual components for the DataMeadow </li></ul><ul><ul><li>More complex visual queries </li...
Questions? <ul><li>Contact information: Niklas Elmqvist E-mail:  [email_address]   WWW: http://www.lri.fr /~elm/ Phone: +3...
Upcoming SlideShare
Loading in …5
×

DataMeadow: A Visual Canvas for Analysis of Large-Scale Multivariate Data

1,152 views

Published on

Supporting visual analytics of multiple large-scale multidimensional datasets requires a high degree of interactivity and user control beyond the conventional challenges of visualizing such datasets. We present the DataMeadow, a visual canvas providing rich interaction for constructing visual queries using graphical set representations called DataRoses. A DataRose is essentially a starplot of selected columns in a dataset displayed as multivariate visualizations with dynamic query sliders integrated into each axis. The purpose of the DataMeadow is to allow users to create advanced visual queries by iteratively selecting and filtering into the multidimensional data. Furthermore, the canvas provides a clear history of the analysis that can be annotated to facilitate dissemination of analytical results to outsiders. Towards this end, the DataMeadow has a direct manipulation interface for selection, filtering, and creation of sets, subsets, and data dependencies using both simple and complex mouse gestures. We have evaluated our system using a qualitative expert review involving two researchers working in the area. Results from this review are favorable for our new method.

Published in: Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
1,152
On SlideShare
0
From Embeds
0
Number of Embeds
7
Actions
Shares
0
Downloads
16
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

DataMeadow: A Visual Canvas for Analysis of Large-Scale Multivariate Data

  1. 1. DataMeadow A Visual Canvas for Analysis of Large-Scale Multivariate Data Niklas Elmqvist (elm@lri.fr) – John Stasko (stasko@cc.gatech.edu) – Philippas Tsigas (tsigas@chalmers.se) IEEE Symposium on Visual Analytics Science & Technology 2007
  2. 2. In media res… <ul><li>Let’s start with a demo! </li></ul>
  3. 3. Parallel Coordinates <ul><li>First proposed by Alfred Inselberg in The Visual Computer in 1985 </li></ul><ul><li>Basic idea : stack dimension axes in parallel, points become polylines </li></ul><ul><li>Advantage : easy to add new dimensions </li></ul><ul><li>As we add dimensions, the parallel coordinate diagram grows horizontally </li></ul><ul><li>Another solution is to transform to polar coordinates and make radial axes </li></ul><ul><ul><li>Starplot diagram </li></ul></ul>
  4. 4. Elements and Dependencies <ul><li>Visual Elements </li></ul><ul><ul><li>Conforms to system-wide data format </li></ul></ul><ul><ul><li>Visual appearance depending on input </li></ul></ul><ul><ul><li>Input and output ports </li></ul></ul><ul><ul><li>Three types : Sources, sinks, transformers </li></ul></ul><ul><li>Dependencies </li></ul><ul><ul><li>Conforms to system-wide data format </li></ul></ul><ul><ul><li>Directed link between two elements </li></ul></ul><ul><ul><li>Propagates data from source to destination </li></ul></ul><ul><ul><li>Interactive updates </li></ul></ul>? … … … …
  5. 5. The DataRose <ul><li>2D starplot display </li></ul><ul><li>Transformer visual element </li></ul><ul><li>Average as black polyline </li></ul><ul><li>Shows data distribution using different representations </li></ul><ul><ul><li>Opacity bands [Fua et al. 1999] </li></ul></ul><ul><ul><li>Color histogram bands </li></ul></ul><ul><ul><li>Parallel coordinates </li></ul></ul>
  6. 6. DataRose Representations <ul><li>Parallel coordinate mode </li></ul><ul><ul><li>See all details </li></ul></ul><ul><ul><li>Cannot see distribution </li></ul></ul><ul><li>Color histogram mode </li></ul><ul><ul><li>LOCS color scale </li></ul></ul><ul><ul><li>Brightness = high value </li></ul></ul>
  7. 7. DataRose Types <ul><li>DataRoses can be of several different types </li></ul><ul><ul><li>Each type represents a specific multi-set operation </li></ul></ul><ul><li>Four types: </li></ul><ul><ul><li>Source : external database loaded from a file </li></ul></ul><ul><ul><li>Union : all input cases combined </li></ul></ul><ul><ul><li>Intersection : input cases that exist in all input sets </li></ul></ul><ul><ul><li>Uniqueness : input cases that exist in one input set </li></ul></ul>
  8. 8. Viewers and Annotations <ul><li>Viewers are sink elements: accept input - no output </li></ul><ul><li>Shows quantitative information for the inputs </li></ul><ul><ul><li>Barchart </li></ul></ul><ul><ul><li>Piecharts </li></ul></ul><ul><ul><li>Histogram </li></ul></ul><ul><li>Annotations support communication of analyses </li></ul><ul><ul><li>Labels </li></ul></ul><ul><ul><li>Notes </li></ul></ul><ul><ul><li>Images </li></ul></ul><ul><ul><li>Reports </li></ul></ul>
  9. 9. Evaluation <ul><li>Evaluation : expert review (think-aloud protocol) </li></ul><ul><li>Participants : two visualization researchers </li></ul><ul><li>Dataset : US Census 2000 </li></ul><ul><ul><li>Three types of open-ended questions: </li></ul></ul><ul><ul><ul><li>Direct facts: “What is the average house value in Georgia?” </li></ul></ul></ul><ul><ul><ul><li>Comprehension: “Which state has the highest ratio of small and expensive houses?” </li></ul></ul></ul><ul><ul><ul><li>Extrapolation: “Is there a relation between fuel type and building size in Alaska?” </li></ul></ul></ul><ul><li>Results : positive, has lead to new design iterations </li></ul><ul><ul><li>Participants were able to solve questions </li></ul></ul><ul><ul><li>Interaction quoted as main benefit </li></ul></ul>
  10. 10. Contributions <ul><li>A highly interactive visual canvas ( DataMeadow ) for multivariate data analysis using multiple small visualization components </li></ul><ul><li>A visual representation ( DataRose ) based on axis-filtered parallel coordinate starplots that can be linked together into interactive visual queries </li></ul><ul><li>Results from a qualitative user study showing the use of our system for multivariate data analysis </li></ul>
  11. 11. Future Work <ul><li>Additional visual components for the DataMeadow </li></ul><ul><ul><li>More complex visual queries </li></ul></ul><ul><ul><li>Additional annotation and communication support </li></ul></ul><ul><li>Non-standard input devices </li></ul><ul><ul><li>Pen-based interfaces </li></ul></ul><ul><li>Non-standard output devices </li></ul><ul><ul><li>Large displays </li></ul></ul><ul><ul><li>Collaborative visual analytics </li></ul></ul>
  12. 12. Questions? <ul><li>Contact information: Niklas Elmqvist E-mail: [email_address] WWW: http://www.lri.fr /~elm/ Phone: +33 1 69 15 61 97 Fax: +33 1 69 15 65 86 </li></ul>Pictures courtesy of Helene Gregerström Taken at the Atlanta Botanical Gardens http://www.aviz.fr/

×