Proposal 12 - Visual Analytics
- 1. “ The use of visual analytics tools for unstructured content analysis” David Whitehead TBA, Vancouver Visual Analytics
- 3. My Proposal Project Investigate adopting visual analytics tools for unstructured content analysis and provide these tools as a service for TBA's and IS's in support of client and sector needs. Stage 1: needs analysis What unstructured content analysis does CISTI perform? What types of client questions could/should we use VA tools to help answer? How would different CISTI teams make use of VA tools? What experience does CISTI have with VA and unstructured content analysis tools? Stage 2: survey available tools What tools are available? What are the tradeoffs for each tool? How well do they meet CISTI's needs? Stage 3: pilot study Select and deploy 1 or 2 VA tools for use in a pilot study Select and train up to 3 IS/TBA teams in the use of Visual Analysis techniques and tools Each team to use VA tools to work with at least 2 client projects Teams to report on the effectiveness, usability, strengths and weaknesses of the selected VA tools. Clients to feedback on the usefulness of the VA tool analysis/findings Deliverables Unstructured content analysis tools needs assessment Survey of available visual analytics tools Train 1 or 2 IS/TBA teams on Visual Analytics techniques Pilot Study report
- 4. My Proposal Project Investigate adopting visual analytics tools for unstructured content analysis and provide these tools as a service for TBA's and IS's in support of client and sector needs. Stage 1: needs analysis What unstructured content analysis does CISTI perform? What types of client questions could/should we use VA tools to help answer? How would different CISTI teams make use of VA tools? What experience does CISTI have with VA and unstructured content analysis tools? Stage 2: survey available tools What tools are available? What are the tradeoffs for each tool? How well do they meet CISTI's needs? Stage 3: pilot study Select and deploy 1 or 2 VA tools for use in a pilot study Select and train up to 3 IS/TBA teams in the use of Visual Analysis techniques and tools Each team to use VA tools to work with at least 2 client projects Teams to report on the effectiveness, usability, strengths and weaknesses of the selected VA tools. Clients to feedback on the usefulness of the VA tool analysis/findings Deliverables Unstructured content analysis tools needs assessment Survey of available visual analytics tools Train 1 or 2 IS/TBA teams on Visual Analytics techniques Pilot Study report OVERLY AMBITIOUS
- 5. Reality One Visual Analytics tool: “ Starlight” from Futurepoint Systems Two projects: Catalog of the BC Wireless Industry Finding an appropriate partner for an enterprise USB Flash drive solution
- 6. I tried one Visual Analytics tool out on two projects: Catalog of the BC Wireless Industry Finding an appropriate partner for an enterprise USB Flash drive solution Reality Practical
- 7. Why Visual Analytics I don’t look for terrorists, but I do reduce risk
- 8. Visual Analytics “ People use visual analytics tools and techniques to synthesize information and derive insight from massive, dynamic, ambiguous, and often conflicting data; detect the expected and discover the unexpected; provide timely, defensible, and understandable assessments; and communicate assessment effectively for action.” From: http://en.wikipedia.org/wiki/Visual_analytics
- 9. Visual Analytics Visual Analytics Sounds a lot like CISTI! synthesize information derive insight from massive, dynamic, ambiguous, conflicting data detect the expected and discover the unexpected provide timely, defensible, and understandable assessments communicate assessment effectively for action
- 10. Why explore Visual Analytics Clients come to IS’s and TBA’s expecting "the magic answers" and "a crystal ball“ Our success depends on our ability to meet that un-written brand promise, as unrealistic as it may be. IS’s and TBA’s need to be the experts in information retrieval and analysis, including sophisticated analysis of unstructured content. IS’s and TBA’s have the expertise to interpret and convey the results to business clients.
- 11. Visual Analytics Turns This Into This 1 - 10 of about 3,660,000 for enterprise usb content distribution
- 12. Demo Incredibly Amazing Demo
- 13. Multiple Views
- 14. Multiple Views
- 15. Subject Views
- 16. Good Clustering of Data Helps identify key topics for analysis Quickly eliminates results of no value Provides some confidence in the uniqueness of a solution Allows broader coverage than manual approaches
- 17. Bad Deeper Analysis requires a lot of preparation and data manipulation Getting the data from here to something consumable is hard Requires a lot of learning to get meaningful results Best used by experts in a domain
- 18. Awesome! Data input tools Automatically crawl sites Convert unstructured documents into data Look for multiple parameters simultaneously Automatically extract entities like people, places and companies
- 19. Test 1 Enterprise USB Flash Application Test Traditional Approach Visual Analytics Approach 7 hours Read 53 web pages Identified 20 potential partners Narrowed to 4 likely partners Identified 2 key competing approaches 2 hours Analyzed about 600 web pages Identified 1 likely partner Identified 7 competing approaches
- 20. Test 2 Wireless Industry Profile Traditional Approach Visual Analytics Approach 1 month Read about 1000 web pages Classified 93 pre-defined technologies across 277 mobile industry firms in 2 cities (Vancouver & Ottawa) 8 hours 1 st web crawler pass Analyzed about 11000 web pages starting with links to 137 companies Identified 123 separate technologies
- 21. General Observations Analytics result in much faster generation of garbage out Many trials needed to get the automation right Once right – automation dramatically helps in large environmental scans. The two techniques complement each other well, help to double check findings Helps find more outliers Helps identify useful items to look for Investigative techniques are essential to drawing real conclusions Very hard to communicate the process and the comparative value of results using the tools alone.
- 22. Visual Analytics “ Visual Analytics is the integration of interactive visualization with analysis techniques to answer a growing range of questions in science, business, and analysis. It can attack certain problems whose size, complexity, and need for closely coupled human and machine analysis may make them otherwise intractable.” From: http://en.wikipedia.org/wiki/Visual_analytics
- 23. Why learn more… Anyone can use Google… VA tools are the tools used by leaders in large scale information analysis such as government security forces Visual Analytics help make sense of information overload Making sense of information overload is the essential skill for information analysts. Information overload is every industries problem VA tools are moving out of the security world into the realm of business and scientific analysis expertise in applying tools like VA to analysis tasks in the fields of business and science is what will set CISTI IS’s and TBA’s apart.