Downloaded 10 times
Uploaded byCliff Lampe
2013 05 20 field_directors
The document discusses the pros and cons of using "big data" from social media for social science research. It outlines how social media interactions can be mined for insights but have challenges regarding access, representativeness, validity, and ethics. While big data is inexpensive and covers large scales, more interdisciplinary work is needed to properly test its issues before replacing surveys. Big data is only just beginning to be applied to social questions.
More Related Content
Similar to 2013 05 20 field_directors
2013 05 20 field_directors
- 1. Computational Social Science: ThePros and Cons of 'Big Data’ Cliff Lampe - School of Information, University of Michigan May 20, 2013 Monday, May 20, 13
- 2. Cliff Lampe School ofInformation - associate professor Social media “Socio-technical systems” Primarily a social scientist Monday, May 20, 13
- 3. Samples of myresearch in this area Effects of participation on Facebook Information cascades on Twitter User collaboration on Wikipedia Discussion patterns in large-scale news sites Coordination in massive online games Information seeking on search engines vs. social media Monday, May 20, 13
- 4. Interactions in socialmedia leave communication traces we can mine to understand social processes. These compete with insights from surveys. Monday, May 20, 13
- 5. Defining “Big Data” “BigData” is a rough categorization, a marketing term, and a paradigm shift. Monday, May 20, 13
- 6. Why “big data”has become a big deal... More devices collecting data More data born digital Easier/cheaper to store Better processors New skills / techniques Insights have proven effective Monday, May 20, 13
- 7.
- 8. “Big Data” startedin the physical sciences Monday, May 20, 13
- 9. Big Data isincreasingly being applied to social science questions Monday, May 20, 13
- 10. What counts as“big”? LHC: .001% of sensors lead to 25 petabytes annually. Wikipedia: 17 terabytes Twitter: ~ 10 GB/day How many observations needed to count as “big”? Monday, May 20, 13
- 11. ‘Big Data’ requiremultiple, interlinked skills and tools. Monday, May 20, 13
- 12.
- 13. Challenges in “BigData” Capture Curation Storage Search Sharing Transfer Analysis Visualization Monday, May 20, 13
- 14. Social Media isoften linked with Big Data because it is the likeliest source for human trace data. Monday, May 20, 13
- 15. What is socialmedia? Monday, May 20, 13
- 16. Common characteristics User generatedcontent Direct user-to-user interaction Bundles of applications More than Facebook and Twitter Monday, May 20, 13
- 17.
- 18.
- 19.
- 20.
- 21.
- 22.
- 23.
- 24.
- 25. Social media covera wide variety of sites. Monday, May 20, 13
- 26. Trends in socialmedia use Monday, May 20, 13
- 27.
- 28.
- 29.
- 30. Social media skill Nearly1 million people join Facebook every week People spend on average 16 hours a month on Facebook There are about 250 million Tweets per day People upload 3000 pictures to Flickr every minute Wikipedia has 17 million articles by 91,000 editors YouTube has 490 million unique visitors per month Google + reached 10 million users in 16 days Monday, May 20, 13
- 31. The social medialandscape is constantly changing. Monday, May 20, 13
- 32. People are increasingly enactingtheir social lives in computer mediated channels. Monday, May 20, 13
- 33. Examples of socialmedia / big data social insights Monday, May 20, 13
- 34.
- 35.
- 36.
- 37.
- 38.
- 39.
- 40. Social media arebeing used to capture big social data. Monday, May 20, 13
- 41. Obtaining social mediatrace data “Scraping” APIs (rare) public datasets Partnerships Monday, May 20, 13
- 42. Scraping - usingsoftware to download and store data from publicly available site. Monday, May 20, 13
- 43.
- 44.
- 45.
- 46.
- 47.
- 48. Issues with socialmedia trace data Access Representing results Representativeness Validity Cross-channel difficulty Appropriate skill sets Ethics Monday, May 20, 13
- 49. Access Data often ownedby private corporations. Need special skills to access. Monday, May 20, 13
- 50. Representing results Probability testingbreaks down. Visualization is common, but limited. Training audiences for new data analysis. Monday, May 20, 13
- 51. Representativeness* How accurately do socialmedia users represent the larger population? How do you rigorously sample from social media? Monday, May 20, 13
- 52. Validity Social media usersare performing (though don’t know scientists are observing them) Different sites have different purposes. Monday, May 20, 13
- 53. Cross-channel use How doyou track one user over multiple social media sites? Monday, May 20, 13
- 54. Appropriate researcher skill sets Combinationof technical and research skills are required. New generation of “data scientists” coming now. Monday, May 20, 13
- 55. Ethics How can auser opt out? More on a panel next session... Monday, May 20, 13
- 56. Challenges in “BigData” Capture Curation Storage Search Sharing Transfer Analysis Visualization Monday, May 20, 13
- 57. “Pros”of big data It’srelatively cheap Massive scale covers some sins It’s inevitable(?) In some cases, it works. Monday, May 20, 13
- 58. Insensitive Borg says... BigData will make surveys obsolete. Monday, May 20, 13
- 59. Humble Suggestions More interdisciplinarywork. Propose and fund work to test these issues. Don’t pretend it isn’t coming OR is a panacea. We’re just at the beginning of the journey. Monday, May 20, 13
- 60. Social media andsurveys project Can social media data ever replace and/or supplement social measurement, especially for official statistics, based on self-reported answers to questions asked of a representative sample? Fred Conrad, Michael Schober, Josh Pasek Monday, May 20, 13
- 61. Thanks! Cliff Lampe [email protected] Twitter: @clifflampe Slideshare:clifflampe Monday, May 20, 13