Tech Talk | Product Studio

Fall 2016
Cornell Tech
The New York Times (Community)

Transcript

1. dispel Community Delia Casa Frances Coronel Chumeng Xu
2. How might we create a safe place for people of all backgrounds to discuss important local, national and global issues?
3. Demographic Makeup of New York Times Commenters
4. “Over the last year The Times has watched readership fall significantly . . . an extremely worrying sign on a growing platform ” Source: The New York Times Innovation Report - March 23, 2014
8. Reader’s View
9. Reader’s View
10. Commenter Editor Q A &
11. Commenters IBM Watson Alchemy API Prompt user to comment with targeting questions Automatically extract key attributes from article and generate question stems Self identify user attributes System Architecture Commenting Platform Content Management System Editor Database of User Attributes Ad Targeting Automated Social Media Publishing Better Article Recommendations
12. IBM Watson Alchemy API APIs Out There Google Cloud Natural Language API
17. Technically Challenging • using database to generate revenue • lots of data points will have to be aware of each other • user attributes must be associated with each article individually • making sure relevant user attributes are automatically selected • safe filtering (what if it’s an article on racist people?)
18. What’s next? • code it out!