Data Kindness on the Internet: Discussing Valuation and Best Practices
•Download as KEY, PDF•
1 like•676 views
Slides for the Birds of a Feather talk led by Christan Grant on Data kindness on the internet.
Recommended
More Related Content
Viewers also liked
Viewers also liked (20)
Similar to Data Kindness on the Internet: Discussing Valuation and Best Practices
Similar to Data Kindness on the Internet: Discussing Valuation and Best Practices (20)
Data Kindness on the Internet: Discussing Valuation and Best Practices
- 1. Data Kindness on the Internet Christan Grant @cegme Christan Grant • Tapia 2011 • Birds of a Feather
- 2. Data Kindness on the Internet BoF Rules: Ask questions, address to the group not just me Goal: Get researchers using (open) data in research and ventures. Last chunk of minutes will be a group bonding activity Christan Grant • Tapia 2011 • Birds of a Feather
- 3. Data Kindness on the Internet • Data Bandwagon • Data Valuation • Dealing with Data • Good and Bad • Open Data Breakout Christan Grant • Tapia 2011 • Birds of a Feather
- 4. People Dr. Tyrone Grandison An ACM Distinguished Engineer, and program manager for IBM core healthcare services. Dr. Grandison is world renowned leader in the area of information privacy and security. Christan Grant • Tapia 2011 • Birds of a Feather
- 5. People Dr. Kun Liu Dr. Liu is researching behavioral targeting issues in the Yahoo! Display advertising team. He has co-authored three book chapters on privacy- preserving social-network analysis and privacy- preserving data mining. Christan Grant • Tapia 2011 • Birds of a Feather
- 6. People Dr. Christan Grant A doctoral student at the University of Florida Database Research Center, has worked on projects to extract information from hidden data bases and collections of online. Christan Grant • Tapia 2011 • Birds of a Feather
- 7. lets talk about data • Web 2.0 is predicated in the fact users will upload their data • Data → Knowledge → Power • A Market for Lemons (George Akerlof) Christan Grant • Tapia 2011 • Birds of a Feather
- 8. Valuation of Data “If you are not paying for something, you are not the customer, you are the product.” Christan Grant • Tapia 2011 • Birds of a Feather
- 9. Valuation of Data “Personal data is the oil of the Internet and the new currency of the digital world.” -- Megan Kuneva, Consumer Commissioner Christan Grant • Tapia 2011 • Birds of a Feather
- 10. Valuation of Data • Facebook • Ad Sales 2010 is $1.21 Billion • 90 “pieces on content” created per month per user (3 per day) Christan Grant • Tapia 2011 • Birds of a Feather
- 11. Valuation of Data Christan Grant • Tapia 2011 • Birds of a Feather
- 12. Valuation of Data Facebook Christan Grant • Tapia 2011 • Birds of a Feather
- 13. Valuation of Data Facebook $1.21 Billion ÷ (90 data/(month x user) x 12 months) Christan Grant • Tapia 2011 • Birds of a Feather
- 14. Valuation of Data Facebook $1.21 Billion ÷ (90 data/(month x user) x 12 months) $1.12 Million per user/data Christan Grant • Tapia 2011 • Birds of a Feather
- 15. Valuation of Data Facebook $1.21 Billion ÷ (90 data/(month x user) x 12 months) $1.12 Million per user/data 500 Millions users!! Why The Face?? Christan Grant • Tapia 2011 • Birds of a Feather
- 16. Valuation of Data Facebook $1.21 Billion ÷ (90 data/(month x user) x 12 months) $1.12 Million per user/data 500 Millions users!! Why The Face?? Note: Inaccurate, we need to know how much ads per user action Christan Grant • Tapia 2011 • Birds of a Feather
- 17. Valuation of Data • Google • ~97% of revenue from ads • $28 Billion from Ads in 2010 Christan Grant • Tapia 2011 • Birds of a Feather
- 18. Valuation of Data • Yahoo! • 90% of revenue from Ads • Revenue $6.460 billion (2009) • 6 Million Users Christan Grant • Tapia 2011 • Birds of a Feather
- 19. Valuation of Data • Color.com New location-based photo sharing “elastic” social network. Initial seed investment of $41 Million dollars Christan Grant • Tapia 2011 • Birds of a Feather
- 20. Valuation of Data • Color.com New location-based photo sharing “elastic” social network. Initial seed investment of $41 Million dollars How?? Bubble?? Christan Grant • Tapia 2011 • Birds of a Feather
- 21. Valuation of Data • Color.com • Answer: Data!! • Grabs more data points per picture, patents for garbing user location and proximity. • “sound levels, Bluetooth readings, light readings, antenna strength, the time - even the direction you're pointing your phone - and more and uses it all to determine your proximity to other users.” - rww • “It is a research company and a data mining company” - Bill Nguyen • “Our data is so accurate that we know where you are” - Bill Nguyen Christan Grant • Tapia 2011 • Birds of a Feather
- 22. Valuation of Data * Data is valuable for people who know how to use it • Part of the value calculation is the cost of a data breach • Buyers are largely advertisers (People say “we shouldn’t put value on data” ... to me that is like saying don’t put value on crude oil) Christan Grant • Tapia 2011 • Birds of a Feather
- 23. Dealing with Data • Categories of data? • How to collect internet oil? • What is data kindness? • Examples Christan Grant • Tapia 2011 • Birds of a Feather
- 24. Dealing with Data • Categories of Data • Geographic • Demographic • Behavioral Christan Grant • Tapia 2011 • Birds of a Feather
- 25. Dealing with Data • Collecting internet oil • Active • Passive Christan Grant • Tapia 2011 • Birds of a Feather
- 26. Dealing with Data • Active data collection • Collecting data by proactively seeking and obtaining • Common Examples: • Scraping websites • Credit card applications • Medical procedure information Christan Grant • Tapia 2011 • Birds of a Feather
- 27. Dealing with Data • Passive data collection • Information is obtained through monitoring streams • Common Examples: • Cookies • Telecom call logs (Just press record) Christan Grant • Tapia 2011 • Birds of a Feather
- 28. Dealing with Data • Data Kindness • Polite method to collect/use/provide/destroy data. Say what you are going to do, do it, and do only this. • “Kind data” transaction model is a two-way gentle person hand-shake. • If a website has a privacy policy, follow it. • API’s/Protocols make this formal Christan Grant • Tapia 2011 • Birds of a Feather
- 29. Dealing with Data • Data Kindness ❿Insurance companies • Active • Insurers have long used blood and urine tests to assess people's health • Instead, Information comes from online and offline activities to predict life expectancies • Information comes from warranty cards • Registering at websites (weight loss tips websites, bungee jumping websites) • Find higher risk information, looking for clues about life style that play into obesity, hypertension and diabetes. Insurers are looking for healthy people, people who exercise a lot. Christan Grant • Tapia 2011 • Birds of a Feather
- 30. Dealing with Data • Data Kindness ❿Credit card companies • Passive • Visa Europe is looking into using your phone location to prevent fraud • This is used to help reduce false positives • By 2015, more than 15% of cards will be validated by mobile location Christan Grant • Tapia 2011 • Birds of a Feather
- 31. Dealing with Data • Data Kindness ❿Street Line Networks • Put sensors in city parking spaces to help people find parking spots. • Active or Passive? Christan Grant • Tapia 2011 • Birds of a Feather
- 32. Dealing with Data • Data (un)Kindness ❿Google Street View • Active • 240K people opted out of being included in google street view. • Houses were blurred... some were not • German court said it was legal. • Fined €100K France for activities Christan Grant • Tapia 2011 • Birds of a Feather
- 33. Dealing with Data • Data (un)Kindness ❿Websites • Passive • first party cookies -- text files that provide useful services (e.g. shopping cart) • 3rd party cookies -- text file with unique id to track websites you visit • beacons -- small programs, cookies on steroids (e.g. a key logger) • flash cookies -- hard to delete, may replace deleted cookies mostly volume setting Christan Grant • Tapia 2011 • Birds of a Feather
- 34. Dealing with Data • Data (un)Kindness ❿Websites • Passive • dictionary.com → 234 • msn.com → 207 • comcast.net → 151 • aol.com → 133 • Some phone apps send your phoneid and other identifiable information to 3rd parties. • A large amount of these apps do not have privacy policies Christan Grant • Tapia 2011 • Birds of a Feather
- 35. Dealing with Data • How to use data • Sharing data can have a huge impact • $60 mil Alzheimer Research → 160 papers, 80 in the pipeline • Publish your data as linked-data to semantically combine with other information • Be data stewards and declare data ownership Christan Grant • Tapia 2011 • Birds of a Feather
- 36. Dealing with Data • Protect sensitive information • Anonymize • High tech breach notification act -- must notify individuals when their health information is breached • 2009 US data breach cost $6.75 Mil • 2009 US detection/escalation $0.26 Mil • 2009 US notification cost $0.5 Mil Christan Grant • Tapia 2011 • Birds of a Feather
- 37. Christan Grant • Tapia 2011 • Birds of a Feather
- 38. Dealing with Data • Obtaining data • Open data sets infochimps.org • dev.twitter.com • developers.facebook.com • Scraper Wiki • developers.nytimes.com • Health Tap data http://h4h.healthtap.com/pages/ data Christan Grant • Tapia 2011 • Birds of a Feather
- 39. Dealing with Data • Tools • Graduate students • Scripting language - python/bash/awk • Google refine • Yahoo pipes • Bing maps • ... Christan Grant • Tapia 2011 • Birds of a Feather
- 40. Dealing with Data • Data Quality • Most data is meh - Restaurant data is 3.5-4 stars • Aggregated/fused data is better information is better • “90% of time is spent modeling data” Christan Grant • Tapia 2011 • Birds of a Feather
- 41. Dealing with Data • Data Quality • Integrating data is difficult (Information Integration is a billion dollar business) • Semantic data helps • Larger and more diverse the data sets, the more difficult • Dates/Unicode is are a headache • Separate teams must clearly define downstream data Christan Grant • Tapia 2011 • Birds of a Feather
- 42. Open Data Breakout • Choose or make up a topic • Get together in groups of 5-6 • Discuss a plan to attack this topic • With ~ 15 minutes left we will start presentations Christan Grant • Tapia 2011 • Birds of a Feather
- 43. Open Data Breakout! • Projects • Vision → How many Facebook pictures on average will it take to create a 3D model of a person • Algos/Statistics →How would you count the number of spam bots on twitter • NLP → How can you use language specific wording to predict user location/demographics • Healthcare → How can you use open health data to improve diagnosis. • Robotics → How can a robot use web or social network data to entertain a human companion • Web/Algos → Web based privacy risk score Christan Grant • Tapia 2011 • Birds of a Feather
- 44. Links • Links to resources are at: http://goo.gl/ TCF3x Christan Grant • Tapia 2011 • Birds of a Feather
- 45. Thank you • Questions?! Christan Grant christangrant.com University of Florida Christan Grant • Tapia 2011 • Birds of a Feather
Editor's Notes
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n
- \n