Linked Data, Big Data and User Science at globo.com

Transcript

1. Ícaro Medeiros icaro.medeiros@gmail.com semantica@corp.globo.com I Encontro de Computação Semântica @UFRJ 11/03/2015 LINKED DATA BIG DATA USER SCIENCE @ globo.com
2. INTERACTION CONTENT USER
3. ( icaro, home_globoesporte, pageview@23:00 ) ( icaro, materia_1, scroll+2min@14:00 ) Signals ( materia_1: [messi, neymar, barcelona] ) content description
4. LINKED DATA (content)
5. Ontologies ‣ 288 classes ‣ Person: 65K ‣ Place: 50K ‣ Athlete: 22K ‣ Politicians: 32K
6. Annotation tool
8. Interface follows the ontology Fields Suggest as you type Triples stored in Virtuoso Automatic entity extraction Fast search in Elastic Search
9. Contextual navigation
10. globo esporte .com
11. globo esporte .com
12. globo esporte .com
13. Automatic page generation
15. Intelligent Search
16. BIG DATA
17. Cluster Stats ‣ 10 machines ‣ 1 TB RAM ‣ 500 TB disk ‣ 338 VCores
18. Signal Capturing
19. Beyond clicks (engagement science) ‣ Attention-based metrics ‣ Scroll ‣ Time spent on page ‣ Dwell time ‣ Social Media Analytics http://labs.yahoo.com/publication/beyond-clicks-dwell-time-for-personalization/
20. Shares are noisy http://time.com/12933/what-you-think-you-know-about-the-web-is-wrong/
21. Scroll http://time.com/12933/what-you-think-you-know-about-the-web-is-wrong/
22. Recommendation ‣ TF-IDF ‣ Collaborative Filtering ‣ Users ‣ Content ‣ Latent Factor Analysis
25. USER SCIENCE for news reading
26. User Modeling (for news reading) ‣ Dynamic pro fi ling ‣ Explicit personal data ‣ Interests (implicit) ‣ Temporal constraints: periodicity
27. Signal Capturing Excelsior Signals
28. Semantic User Modeling ‣ Annotations from engaged content ‣ Pro fi le can answer: ‣ My favourite team ‣ City I live in ‣ My hometown
29. Spreading Activation
30. My profile
31. City/State I live in
32. Hometown and State
33. Football team test (3.5MM users) 82%precision 95%precision@top3 * When the user has read at least one article that cites their team
34. How fast? mean request time between interaction and profile update 5 min 48 ms
35. Potential uses ‣ Personalized homepages ‣ Targeted advertising ‣ Granular user/content description ‣ Semantic Recommendation ‣ Clustering ‣ Demographic data ‣ Informed product creation/evolution
36. github.com/ globocom/ IWantToWorkAtGloboCom
37. Ícaro Medeiros icaro.medeiros@gmail.com Semantic team semantica@corp.globo.com globo.com slides icaromedeiros.com.br slideshare.net/icaromedeiros