WUD2008 - The Numbers Revolution and its Effect on the Web
•Download as PPT, PDF•
2 likes•459 views
The document discusses how the "numbers revolution" is affecting the web and user experience design through increased data collection and analysis. It covers how more data availability and analysis tools are enabling new types of applications for decision support, personalization, prediction and visualization. This is changing how people access and think about information by augmenting human cognition with computer analysis. The document provides many examples of current and emerging applications that utilize these approaches in areas like business, health, sports and media.
Report
Share
WUD2008 - The Numbers Revolution and its Effect on the Web
- 1. The Numbers Revolution and its effect on the Web Rich Miller Research Scientist, LexisNexis World Usability Day – 11/13/2008
- 2. What we will cover What the numbers revolution is and how it came about How it is affecting the world and the web How it is changing both the user experience and design approaches How it is being manifested in tools and applications
- 3. The numbers revolution – what is it? Using data and statistics to better see and affect reality Revolutionary impact due to the maturation of the web and its enabling technologies More data and metadata available – more sensing/measuring Faster networks More powerful displays User interface rendering technologies – e.g. Flash Involves… Capturing and organizing data, often large amounts Crunching data to make predictions Slicing, dicing and visualizing data to aid decision-making and discovery Information overload revisited More information (in the form of metadata) needed to enable the consumption of large amounts of data
- 4. Process and related domains
- 5. Hans Rosling and trendalyzer http:// www.youtube.com/watch?v =hVimVzgtD6w Start at 2:15, goto 5:15 Sequel introducing predictive analytics demo
- 6. How is it changing how people think? Focus thinking on what humans do well and let the computers handle the tough thinking jobs Less guessing and hypothesis-testing, more discovery Predicting based on data and sometimes relegating expertise to asking the right questions Analytics-powered thinking essentially amplifies the main advantage of the internet: information at your fingertips there is no reason that both information and analysis should not be at a user's fingertips
- 7. Where is it having the most impact? Organization decision-making and strategy Includes business intelligence and CRM Predictive modeling Inform general product/UI design and usability User behavior analytics Navigation Consumption Personalization Feedback and testing e.g. randomized web tests Enable more powerful, analytics-fueled applications Decision-support tools Recommendation systems Interactive visual interfaces to enhance… Consuming content Socializing and collaborating
- 8. Importance of the metadata layer As the amount of data available through the web is continuing to increase, a new layer of metadata is being created around it The metadata layer yields multiple benefits, e.g. framework for organizing data Fuel for UIs personalization individual user productivity and decision-support
- 10. Moneyball , Baseball Abstract Sports geeks as pioneers Evaluating talent Determining strategy 1980s - Bill James Challenged conventional wisdom 1990s – Fantasy Sports The Lonious Monks 2000s – Major League Baseball Moneyball – Oakland A’s Red Sox hire Epstein, James Today – part of the game NBA – e.g. Houston Rockets Science
- 11. Freakonomics Applying economic analysis to understand human behavior Like Bill James, Stephen Levitt challenges conventional wisdom Investigations include… Why do teachers and sumo wrestlers cheat? Why should you be suspicious of your real-estate agent? Why do most crack dealers live with their mothers? Is there a link between abortion and crime rate? Why does good parenting not really affect educational performance? How is a child’s name affected by his/her socioeconomic position? Why it is safer to own a gun than a swimming pool? Freakonomics blog
- 12. Super Crunchers , Ian Ayres Ayres = Yale law professor Collaborates with Freakonomics guys Focus on very large datasets and multiple regression analysis Goes beyond BI by combining predictive models with randomized experiments How statistical evidence can supplement/replace human judgement. exposes experts’ limitations in predicting seeking decision-making help from computers should be normal part of business. Topics include medicine, education, business, sports, and winemaking.
- 13. The Wisdom of Crowds In this case, the numbers are the people in the crowd…the more the better Under certain conditions, the crowd’s decisions are superior to the individual’s 4 conditions necessary diversity of opinion independence of opinions decentralization of power aggregation into a group answer Again, experts exposed as inferior – to crowd in predicting … and “less-bright” folks are essential! Examples include auto traffic behavior disease tracking and treating navigating the internet
- 14. Competing on Analytics How leading companies collect, analyze, and act on data It takes an investment, a plan,and the discipline to stick to it. Applications include… Supply chain, customer relations, pricing, HR, product quality, R&D. Companies include Marriot International Capital One Oakland A’s, Boston Red Sox Procter & Gamble Bottom line Make it a normal part of your business
- 15. WIRED: The End of Science “ All models are wrong, but some are useful” – George Box (statistician) “ All models are wrong, and you can do without them” – Peter Norvig (Google) A shift from traditional scientific models to ones based on collecting and analyzing large bodies of data – “the petabyte age” essentially discovering the truth as opposed to making predictions and testing them. Example applications include disease surveillance, farming, physics, legal discovery, news/event monitoring, astronomy, archaeology, airfare analysis, politics/elections, web data management, terrorism insurance. Counterpoint: “If the data wasn't collected to fit the model, the model is probably wrong” – Mark Wasson (LexisNexis) Also, we need a better measure of database size “ information objects” rather than storage space
- 16. Related (but unread by presenter) Similar Wikinomics : How Mass Collaboration Changes Everything The Long Tail : Why the Future of Business is Selling Less of More Differing perspectives Nassim Nicholas Taleb The Black Swan : The Impact of the Highly Improbable Fooled by Randomness : The Hidden Role of Chance in Life and in the Markets
- 17. How is it affecting business/web? Using numbers to get smarter is quickly becoming a competency necessary to effectively compete Numbers used for improving both… Internal decision-making Quality of products for customers Provides fuel for, and often demands existence of, visualization Giving rise to new wave of tools and apps Business intelligence Charting and social/sharing/collaboration Decision-support Information-seeking and research
- 18. Implications for product/UI design Opportunities bridging the gap between normal users and complex statistical analysis tools allowing human brain to do what is does best and let the computer play an augmenting role Creating compelling visual information spaces to explore Challenges adding power while preserving UI simplicity Screen space allocation and management avoiding misleading conclusions or data views…pitfalls include small sample sizes inappropriate mapping of data to charts managing data clean-up issues data normalization – consistent label per entity label abbreviation – so they can fit into charts facilitate learning by integrating new tools with familiar tools
- 19. Best answer vs. best picture The information-seeking UX lies along a continuum from a single answer to an information space Continuum anchored by different views/interfaces Providing best answer – crunching and analyzing data and coming up with the best answer e.g. how old is a person expected to live? Providing a view – organizing and rendering data in a way for users to best understand it and use it. e.g. voting patterns by county in Ohio The continuum reflects the degree of interaction a user has with an application i.e. the broader the answer space, the more the user needs control over manipulating a view of the space
- 20. UI-related technologies/approaches Advances in visualization and infographics 2 fields merging as computer graphics improve New York Times setting the pace Rich internet applications (RIAs) Provide virtually unlimited rendering of graphics and visualization Large displays and new pointing devices
- 22. BI visualization and dashboards Example = Tableau Desktop
- 23. Ian Ayres prediction tools Featured in Super Crunchers Predict the Value of Bordeaux (Ayres - may require free JAVA download ) Personal/Family Predict how long your marriage will last (Political Calculations) Predict Your Child's Adult Height (University of Saskatchewan) Consumer Applications Predict the Market Value of Your Home (Zillow) Fun/Sports (and gambling?) Predict NFL game (NFL Picker) Politics/Government Predict Where Your Tax Money Is Spent (Tax Break Down) Media Predict Demographics of Who Will Use a Webpage (Microsoft adCenter) Health Predict Your Liklihood of Illness or Disability (Northwestern Mutual) Money/Business Predict whether a publicly traded company will file for bankruptcy (Political Calculations) Macroeconomic Predict the odds of a U.S. recession in the next 12 months (Political Calculations)
- 24. Predictive Analytics - SAS
- 25. Small-scale decision-support - visual i|o Should the pitcher be replaced?
- 26. Fantasy sports decision-support CBS Sportsline Fantasy Basketball The outcome of my decisions Who should I start? Who should I add to my team? What moves are others making?
- 27. Infographics - Catalogtree.net Candidate for a geo view
- 28. Infographics – weighted electoral map
- 29. Infographics-visualization convergence New York Times
- 30. Hans Rosling and trendalyzer http:// www.youtube.com/watch?v =hVimVzgtD6w Start at 2:15, goto 5:15 sequel demo
- 31. Info space browsers - Bestiario
- 32. Data+viz sharing - Many-eyes Powers New York Times Visualization Lab example
- 33. Data+viz sharing - Swivel
- 34. Data+viz sharing - lastfm
- 35. Professional info-seeking - LexisNexis LNIS - volume of European news coverage of US Banks over time Total Patent # of patents over time by authority/country Courtlink Strategic Profiles - types of matters handled by law firm
- 36. Pro info-seeking - ILOG/Elixir CIA World Factbook
- 37. Consumer info-seeking – Stamen viz trulia See document: trulia oreilly post on trulia map See document: stamens-map- for.html adobe tour of california RIA See document: adobe digg labs See document: digg global business network See document: gbn root markets See document: root
- 38. What about usability in transportation? A ripe area for using numbers to better view and affect reality Applications Traffic analysis – optimizing traffic efficiency Public transportation planning Trip planning Monitoring vessels/vehicles – intelisea Geo-maps with info layers GPS location trackers – e.g. where is the family car?
- 39. Recommendations for UX pros Acknowledge and discover your newly expanded solution space Seek out and use number-fueled apps Understand user cognitive limits … and provide tools for tasks they need help with View visualization design as an extension of UI design Get a handle on data cleanup and formatting Embrace RIAs and the tools to make them
- 40. Predictions for 2009 and beyond Mining and analyzing data for intelligence becomes common practice for more companies RIA technologies gain momentum Visualization continues to proliferate The analytics shakeout begins as users discover which numbers are not as useful Broken Social Scene – social apps will get hit the hardest Decision-support apps hit the mainstream Increased emphasis on data cleanup Companies with nimble data better compete The world becomes smarter, more objective
- 41. Questions and related resources If we run out of time for questions, feel free to contact me… in person later today [email_address] [email_address] Previous public talks WUD 2007 talk on web applications MVCS talk on Web 2.0 SOASIS talk on Visual Thinking Miami-IMS talk on Trends in HCI