SyrtAPI is a new entertainment platform that combines music and book data from multiple sources like MusicBrainz and NYTimes reviews. It uses Linked Data and SPARQL queries to extract lyrics and reviews and recommend songs that match the content of the books. The team learned about using Linked Data vocabularies and linking datasets while building the prototype, which currently retrieves 25 lyrics and 10 book reviews through its pipeline. Future work includes adding more data sources, developing a mobile app, and using natural language processing to better analyze texts.

1. Surtaki - Syrtaki
Group n. 7: Ignacio Traverso, Elena González-Blanco,
Sebastien Ballesteros, Abduladem Aljamel

2. The idea…
Many people like listening to music while they read
books…

3. The idea…
But sometimes is not to easy to match your favorite music
with the content of the book you are reading!!!

4. But we offer a

5. What is SyrtAPI?
It’s a new way of entertainment

6. What is SyrtAPI?
It’s a new way of entertainment
It saves time

7. What is SyrtAPI?
 It is a linked data based application
 It combines different datasets
 www.musicbrainz.org
 http://www.geonames.org
 http://dbpedia.org

8. What is SyrtAPI?
 With cool stuff!
 Music lyrics
http://lyrics.wikia.com
 And book reviews
NYTimes Reviews

9. How to cope with this …??
Syrtaki team, 4 september 2014, 16:00 …

10. But finally inspiration arrived…
Syrtaki team, 4 september 2014, 19:00 …

11. SPARQL Musicbrainz Dataset

12. SyrtAPI Architecture
Wikilyrics
HTML
NYTimes
Bestseller
JSON
NLP Module
CONVERTER
JSON LD
Recommended
URI songs
Music Brainz
(RDF)
BoilerPipe
Lyric
Extranctor
Harvester
SPARQL
COUCH – DB
(JSON)
JS
DOM
Client #LD
JS
Client #LD
JS
Client #LD
JS

13. Conclusions & Results
 Always bet on JavaScript & JSON LD
 Cool stuff like Linked Data Fragments
 Full SPARQL query client side is 6 lines of JS!
 It’s good to combine classical restful API,
hypermedia and Linked Data

14. Conclusions & Results
 So far we got
 Pipeline ready waiting for internet connection!
 25 lyrics
 10 books reviews
 a cool website

16. Lessons Learned
 We’ve learned a lot! … not only SPARQL, but also
 How to use Linked Data vocabularies
 How to link Linked Data datasets
 The difficulty of finding data
 Teamworking
 Other Cool technologies

18. Future work…
 Adding more datasets
 Adding more and different sources of book reviews
 Developing a mobile app html 5
 Link our app to music libraries like iTunes
 Add more NLP to anaylse the texts
 Make money