BOTTARI: How to offer innovative services listening to the Social Media

1,026 views

Published on

The presentation about BOTTARI I gave at Senso Comune workshop

Published in: Business, Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
1,026
On SlideShare
0
From Embeds
0
Number of Embeds
186
Actions
Shares
0
Downloads
6
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

BOTTARI: How to offer innovative services listening to the Social Media

  1. 1. BOTTARI: How to offer innovative services listening to the Social Media Emanuele Della Valle DEI - Politecnico di Milano emanuele.dellavalle@polimi.it http://emanueledellavalle.org
  2. 2. • Introduction • Architecture • Components • Query Processing • Evaluation – Efficacy – Efficiency • Conclusions Agenda 227.02.2012 - Workshop di Senso Comune, Trento, Italy
  3. 3. 327.02.2012 - Workshop di Senso Comune, Trento, Italy Introduction A stream of live opinions • Micro-posts carry the live opinion of social media users about topics and points of interest. 27.02.2012 - Workshop di Senso Comune, Trento, Italy
  4. 4. • In 2011, an average of three million tweets per day was posted in Seoul. • A large number of these micro-posts are about restaurants, bars, coffees and many other semi-public points of interest (POIs) in the city. • Let us consider, Insadong district with its +200 restaurants Introduction Seoul’s stream of live opinions 427.02.2012 - Workshop di Senso Comune, Trento, Italy Number of tweets per day related to Insadong’s restaurants Insadong’s restaurants ordered by decreasing frequency of tweets
  5. 5. • Trusting this collective opinion, can novel commercial and social services be considered? • The case of BOTTARI Introduction A commercial/social opportunity? 527.02.2012 - Workshop di Senso Comune, Trento, Italy http://www.youtube.com/watch?v=XGOKe_lhSks
  6. 6. • An augmented reality application for personalized recommendation of restaurants in Insadong (Seoul) Introduction What have you seen? 627.02.2012 - Workshop di Senso Comune, Trento, Italy
  7. 7. • Yes and no! • Same use case • BOTTARI listens to the social media Introduction Yet another ? 727.02.2012 - Workshop di Senso Comune, Trento, Italy
  8. 8. outout Query RewriterQuery Rewriter Query Evaluator Query Evaluator Machine LearningMachine Learning Streaming Processor Streaming Processor Geo Filtering Geo Filtering Geo- spatial KB Geo- spatial KB Social Media Analyzer Social Media Analyzer Architecture Android Client 27.02.2012 - Workshop di Senso Comune, Trento, Italy 8
  9. 9. outout Query RewriterQuery Rewriter Query Evaluator Query Evaluator Machine LearningMachine Learning Semantic Stream Processor Semantic Stream Processor Geo Filtering Geo Filtering Geo- spatial KB Social Media Analyzer Social Media Analyzer Architecture Geo-Spatial KB 27.02.2012 - Workshop di Senso Comune, Trento, Italy 9
  10. 10. outout Query RewriterQuery Rewriter Query Evaluator Query Evaluator Machine LearningMachine Learning Semantic Stream Processor Geo Filtering Geo Filtering Geo- spatial KB Geo- spatial KB Social Media Analyzer PUSH: Data Initiated Architecture PUSH Segment – Data Initiated 27.02.2012 - Workshop di Senso Comune, Trento, Italy 10
  11. 11. out Query Rewriter Query Evaluator Machine Learning Semantic Stream Processor Semantic Stream Processor Geo Filtering Geo- spatial KB Geo- spatial KB Social Media Analyzer Social Media Analyzer PULL:QueryInitiated Architecture PULL Segment – Query Initiated 27.02.2012 - Workshop di Senso Comune, Trento, Italy 11
  12. 12. out Query Rewriter Query Evaluator Machine Learning Semantic Stream Processor Geo Filtering Geo- spatial KB Social Media Analyzer PULL:QueryInitiated PUSH: Data Initiated Architecture The Entire BOTTARI System 27.02.2012 - Workshop di Senso Comune, Trento, Italy 12
  13. 13. geo:SpatialThing sioc:UserAccount sioc:id(xsd:string) geo:NamedPlace twd:Tweet twd:messageID(xsd:string) twd:messageTimeStamp(xsd:string) sioc:creator_of twd:talksAbout twd:reply twd:retweet sioc:has_creator twd:talksAboutNeutrally twd:talksAboutPositively twd:followertwd:following twd:post sioc:Post sioc:content(xsd:string) twd:TwitterUser twd:screenName(xsd:string) twd:discuss twd:talksAboutNegatively Components SOR - Geo-Spatial KB 1327.02.2012 - Workshop di Senso Comune, Trento, Italy geo:NamedPlace twd:Tweet twd:messageID(xsd:string) twd:messageTimeStamp(xsd:string) twd:talksAbout twd:reply twd:retweet ator twd:talksAboutNeutrally twd:talksAboutPositively twd:post twd:discuss twd:talksAboutNegatively
  14. 14. Components Social Media Analyzer 1427.02.2012 - Workshop di Senso Comune, Trento, Italy Micro-post Morphologically Analyzable? Rule based Analysis Auto generated rules Auto generated rules Learned documents SVMs Syllable KernelSyllable Kernel Sentiment of the tweet Yes No • Precision tests: – Auto-generated rules ≈ 70% – Manually-coded rules ≈ 90% – Syllable kernel ≈ 50~60% • Our target > 85%
  15. 15. +1 for POI Count +1 for POI Q [1 GRAPH]Most Liked POIs Q Top-10 POIs 1. 2. 3. [7 GRAPHs] Weekly Plot [31 GRAPHs] Monthly PLot [1 DAY] [1 DAY] Sum +1 for POI Q [7 DAYs] Sum +1 for POI Q [31 DAYs] [1 GRAPH] [1 GRAPH] Weekly Map Monthly Map Re-stream Analyze Publish Visualize 1. 2. 3. Re-stream from database C-SPARQL Query Windower List View Plot Line View Heatmap View LEGEND Components Semantic Stream Processor 1527.02.2012 - Workshop di Senso Comune, Trento, Italy
  16. 16. • SUNS (Statistical Unit Node Sets) – A machine learning framework for inductive materialization • Detects interesting data patterns • Predics RDF-triples – i.e., which restaurant a user will tweet positively about – Caractheristics • Capability to deal with sparse, high-dimensional and incomplete data • Multivariate latent space based approach • Modularized approach for easily integrating contextual information Components Machine Learning Component 1627.02.2012 - Workshop di Senso Comune, Trento, Italy
  17. 17. out Query Rewriter Query Evaluator Machine Learning Semantic Stream Processor Geo Filtering Geo- spatial KB Social Media Analyzer PULL:QueryInitiated PUSH: Data Initiated Query Processing An Ontology-based Approach 27.02.2012 - Workshop di Senso 17
  18. 18. SELECT DISTINCT ?poi ?name ?lat ?long ?numPos ?prob WHERE { ?poi a ns:NamedPlace ; ns:name ?name ; geo:lat ?lat ; geo:long ?long . FILTER (f:within_distance(37.5, 126.9, ?lat, ?long, 200)) FILTER (f:dest_point_viewing(37.5, 126.9, ?lat, ?long, 90, 200)) { :someUser sioc:creator_of ?tweet . ?tweet twd:talksAboutPositively ?poi . WITH PROBABILITY ?prob ENSURE PROBABILITY [0.5..1) } ?poi twd:numberOfPositiveTweets ?numPos . } ORDER BY DESC(?numPos), ?prob, f:distance(37.5, 126.9, ?lat, ?long) LIMIT 10 1827.02.2012 - Workshop di Senso Comune, Trento, Italy GEO-SPATIAL FILTER RECCOMENDENDATION REPUTATION Query Processing An Example Query
  19. 19. outQuery Evaluator Query Rewriter Machine Learning Semantic Stream Processor Geo Filtering Geo- spatial KB Social Media Analyzer The issued query is rewritten in three queries; one for each run time component PULL:QueryInitiated PUSH: Data Initiated Query Processing The Input Received is Rewritten 27.02.2012 - Workshop di Senso Comune, Trento, Italy 19
  20. 20. outQuery Evaluator Query Rewriter Machine Learning Semantic Stream Processor Geo Filtering Geo- spatial KB Social Media Analyzer Generates personalized recommendations PULL:QueryInitiated PUSH: Data Initiated Query Processing Each Component Answers 27.02.2012 - Workshop di Senso Comune, Trento, Italy 20
  21. 21. outQuery Evaluator HTTP Query Rewriter Machine Learning Semantic Stream Processor Geo Filtering Geo- spatial KB Social Media Analyzer Gets trends in restaurants’ social reputation PULL:QueryInitiated PUSH: Data Initiated Query Processing Each Component Answers 27.02.2012 - Workshop di Senso Comune, Trento, Italy 21
  22. 22. outQuery Evaluator Query Rewriter Machine Learning Semantic Stream Processor Geo Filtering Geo- spatial KB Social Media Analyzer Gets restaurants close to user location PULL:QueryInitiated PUSH: Data Initiated Query Processing Each Component Answers 27.02.2012 - Workshop di Senso Comune, Trento, Italy 22
  23. 23. outQuery Evaluator Joins results and returns them Query Rewriter Machine Learning Semantic Stream Processor Geo Filtering Geo- spatial KB Social Media Analyzer PULL:QueryInitiated PUSH: Data Initiated Query Processing Results Are Joined and Returned 27.02.2012 - Workshop di Senso 23
  24. 24. Evaluation Efficacy – High Accuracy 2427.02.2012 - Workshop di Senso Comune, Trento, Italy
  25. 25. Evaluation Efficacy – Remembering 90 Days is Enough 2527.02.2012 - Workshop di Senso Comune, Trento, Italy
  26. 26. Evaluation Efficiency – High Throughput 2627.02.2012 - Workshop di Senso Comune, Trento, Italy
  27. 27. • End-user application • Attractive and functional interface • Real-world dynamic data • Fully based on Semantic Web technologies – RDF as common data format between heterogeneous components – SPARQL as query language • Rigorously evaluated – Effective – High throughput for handling dynamic data • Commercial Potential Conclusions 2727.02.2012 - Workshop di Senso Comune, Trento, Italy
  28. 28. • Tony Lee, Seonho Kim (SALTLUX) – Android Application – Social Media Analyzer – Tweets and restaurant descriptiond for Insadong • Marco Balduini (PoliMi) – Semantic Stream Processor • Irene Celino, Daniele Dell’Aglio (CEFRIEL) – LarKC workflows – Ontology • Volker Tresp, Yi Huang (SIEMENS) – Machine Learning component Credits BOTTARI is a joint work with 2827.02.2012 - Workshop di Senso Comune, Trento, Italy
  29. 29. Thank You! Questions? 2927.02.2012 - Workshop di Senso Comune, Trento, Italy Much More to Come! Keep an eye on http://www.streamreasoning.org
  30. 30. BOTTARI: How to offer innovative services listening to the Social Media Emanuele Della Valle DEI - Politecnico di Milano emanuele.dellavalle@polimi.it http://emanueledellavalle.org

×