1 / 32

Viewing Changes Monitoring with Treemaps

Viewing Changes Monitoring with Treemaps. Steve Betten Catherine Plaisant Ben Bederson. SmartMoney.com. SmartMoney.com. SmartMoney.com. A different emphasis. SpaceTree. Treemap. 180 Pages on a Web Server (1 week). brightness = number of requests, no size coding

thelmad
Télécharger la présentation

Viewing Changes Monitoring with Treemaps

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Viewing ChangesMonitoring with Treemaps Steve Betten Catherine Plaisant Ben Bederson

  2. SmartMoney.com

  3. SmartMoney.com

  4. SmartMoney.com

  5. A different emphasis SpaceTree Treemap

  6. 180 Pages on a Web Server (1 week) brightness = number of requests, no size coding i.e. layout stable as long as no page is added/removed

  7. Treemaps for Monitoring • Selected application: web site monitoring • Monitoring: • Look for patterns in web page requests over consecutive time periods • Attribute value changes • Simplification: only one attribute (mapped to color) • Look for web pages that were created or deleted during those weeks • Topology changes • What are our options?

  8. “Value” “Difference of values” Choice of attribute green = increase; red = decrease (brighter = more change) brighter = more requests • Might be more intuitive • but requires to look at 2 values • Change is immediately visible • OK for 2 periods only • Not clear what happen when users need to integrate colors changes over several periods.

  9. “All-at-once” “Slider” Time vs. Space Multiplexing • More pixels per page • Generalizes to any number of periods • Less need to interact • All data available on one screen • Fewer pixels per webpage

  10. All-at-once Slider Main Experiment Value Difference of values

  11. “Show only when existing” Leads to layout shift Intuitive: does not show pages that did not exist! “Always show” Always show pages (colored black when non existing) Precludes using size coding In both cases, small icons can show location of changes Additional small study looking at Page Creation/Deletion

  12. Procedure • 12 subjects (CS students) • Each subject tested each version • Training • 10 min Treemap + 10 min each version • Eight timed tasks • Total duration: about 1 hour 30 minutes • Computer generated datasets • Most pages have small changes • A small number of pages have significant changes

  13. Tasks • Determine if all the pages in a directory grew in popularity (2 tasks) • Find pages with specific patterns of requests e.g. up-down-up(2 tasks) • Find deleted pages • Find created pages and determine if they were popular the first time they existed • Determine if a specific page grew in popularity every week (with no pages added i.e. no possible layout shifts) • Same as 5 with some pages added ( i.e. possible layout shift)

  14. For many questions, users perform equally well using any combination When searching for patterns (e.g. up-down-up) “Difference of values” faster than “value” “Slider” faster than “all-at-once” In tasks where pages were created/deleted “Always show” faster than “show only when existing” Subjects strongly disliked layout shifts Results

  15. For many questions, users perform equally well using any combination When searching for patterns (e.g. up-down-up) “Difference of values” faster than “value” “Slider” faster than “all-at-once” In tasks where pages were created/deleted “Always show” faster than “show only when existing” Subjects strongly disliked layout shifts Results

  16. All-at-once Slider Main Experiment Value Difference of values

  17. Value and Slider (Week 4)

  18. For many questions, users perform equally well using any combination When searching for patterns (e.g. up-down-up) “Difference of values” faster than “value” “Slider” faster than “all-at-once” In tasks where pages were created/deleted “Always show” faster than “show only when existing” Subjects strongly disliked layout shifts Results

  19. For many questions, users perform equally well using any combination When searching for patterns (e.g. up-down-up) “Difference of values” faster than “value” “Slider” faster than “all-at-once” In tasks where pages were created/deleted “Always show” faster than “show only when existing” Subjects strongly disliked layout shifts Results

  20. All-at-once Slider Main Experiment Value Difference of values

  21. For many questions, users perform equally well using any combination When searching for patterns (e.g. up-down-up) “Difference of values” faster than “value” “Slider” faster than “all-at-once” In tasks where pages were created/deleted “Always show” faster than “show only when existing” Subjects strongly disliked layout shifts Results

  22. Disliked the layout shifts In both cases users mostly used the small icons to see location of changes Additional small study looking at Page Creation/Deletion

  23. Subjective Preference • “Difference of values” preferred over “value” • Saves a step (visually determining the difference) • Easier to see hue shifts than intensity changes • “All-at-once” v. “slider” • No difference • Same pattern of preferences for “ease of learning” • Subjects only used details table to check their guesses based on color alone • Very few errors • Only half of subjects adjusted the time interval in the “difference in values and slider” version • User quote: “Different versions are suited to different tasks

  24. Conclusions • Treemaps can be effective for monitoring • Consider using: • “difference of values” instead of “value” • “slider” with an option showing “all-at-once” • Avoid size changes • To download treemap3.2: www.cs.umd.edu/hcil/treemap3(future versions will include monitoring capabilities described today) • Partial support from Chevron-Texaco • Participation of B. Shneiderman and J-D Fekete

More Related