creating the thomson reuters knowledge graph and open permid - odi summit 2015
TRANSCRIPT
![Page 1: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/1.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 1/27
ODI SUMMIT
3 NOVEMBER 2015
@DrDanielASmith
DR. DANIEL A. SMITH, SENIOR DEVELOPER, CORPORATE TECHNOLOGY
CREATING THE THOMSON REUTERSKNOWLEDGE GRAPH AND OPEN PERMID
![Page 2: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/2.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 2/27
, -,
ABOUT THOMSON REUTERSFINANCIAL & RISK
INTELLECTUAL PROPERTY & SC
LEGAL
Comprehensive IP & scientifidecision support tools & servigovernments, academia, publicorporations & law firms.
Critical information, decision software & services to legal, ibusiness and government prof
Critical news, information & analytics,enables transactions, and connectstrading, investing, financial and corporateprofessionals.
TAX & ACCOUNTING
Integrated tax compliance and accountinginformation, software & services forprofessionals in accounting firms,corporations, law firms and government.
REUTERS NEWS
Powered by more than 2,800 journalists reporting in 20languages from bureaux around the world, Reuters isthe world’s largest international news organisation.
![Page 3: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/3.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 3/27
, -,
ABOUT THOMSON REUTERS
![Page 4: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/4.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 4/27
, -,
ABOUT THOMSON REUTERS
![Page 5: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/5.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 5/27
, -,
ABOUT THOMSON REUTERS
• Due to growth by acqusition, weare working with siloed data
• Segregation of content bybusiness domain
![Page 6: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/6.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 6/27
, -,
ABOUT THOMSON REUTERS
• Benefits: Designed, contentcontrolled, edited and publishedby each business
![Page 7: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/7.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 7/27
, -,
CUSTOMER DATA USE
![Page 8: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/8.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 8/27
, -,
CUSTOMER DATA USE
![Page 9: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/9.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 9/27
, -,
KNOWLEDGE GRAPH
![Page 10: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/10.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 10/27
, -,
KNOWLEDGE GRAPH
Company Thomson Reutersname
primaryQuoteQuote
RIC
1977-12-28incorporated
http://tr.comwebsite
ticker
exchange
![Page 11: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/11.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 11/27
, -,
KNOWLEDGE GRAPH
Thomson Reuters name
granted
nameEikon
US20140173400A1 application
2013-10-23 filed
2014-06-19published
uses
Patent
makesProduct
Company
![Page 12: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/12.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 12/27
, -,
KNOWLEDGE GRAPH
Thomson Reuters name
granted
nameEikon
US20140173400A1 application
2013-10-23filed
2014-06-19published
uses
Patent
makesProduct
Thomson Reutersname
primaryQuoteQuote
1977 -12-28incorporated
http://tr.comwebsite
exch
![Page 13: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/13.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 13/27
, -,
KNOWLEDGE GRAPH
![Page 14: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/14.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 14/27
, -,
IDENTITY CHALLENGES
• Entities can have multiple identifiers
• e.g. Organisations have IDs all areas:• Finance and Risk• Tax and Accounting• Legal• IP and Science• News
![Page 15: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/15.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 15/27
, -,
ORGANISATION IDENTIFIERS IN FINANCE AND • MXID
• NDGSymbol• DBSTicker• SDCCusip• SDCID• SEDARIssuer• EdCoID• VEFirmID
• VentureEconomicsID• TMTCompanyID• CIK• DisclosureID• EedbID• GemAlphaNumericID
• RegistrationNumber
• DunsNumber• SinotrustNumber• DatastreamFiId• RegulatoryId• Cusip6• TaxId• RcpId
• EfxId• EjvExchangeCode• Lei• DataStreamId• AllCode• InvestextId
![Page 16: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/16.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 16/27
![Page 17: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/17.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 17/27
, -,
PERMID AS A USEFUL COMMON REFERENCE PO
• Specifically maintained for identity reference & not as a side-effect- Use / context independent – focus is on getting community support & network mass- Unambiguous, consistent interface, doesn’t need interpretation- Well-described & maintained relative to the real world- Stable meaning, persistent, temporal- Coverage & granularity reflect community needs
- Dependable support over time• Everyone knows that everyone else can freely access and use it
- Open licensed- Known quantity to plan against- Creates a network effect
![Page 18: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/18.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 18/27
, -,
TECHNOLOGY STACK
• Content Marketplace• Data Item Registry (i.e., ISO/IEC 11179)• XML
• Knowledge Graph• Semantic Web
• RDF, OWL, SPARQL, SPIN, Jena, Sesame• Big Data
• Apache Big Data Ecosystem - Hadoop, Spark, Kafka, Oozie, Cassandra, Elastic Sea
![Page 19: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/19.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 19/27
, -,
BUILDING THE KNOWLEDGE GRAPH
• The Content Marketplace work gave us the linkage through PermIDs
• Semantic Web and Big Data technologies give a strong starting point to buildinknowledge graph
• Take those technologies and scale them to:• Query or manipulate at scale• Provide lots of data and lots of perspectives on data
![Page 20: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/20.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 20/27
, -,
BUILDING THE KNOWLEDGE GRAPH
• Build a minimal set of tools to put and get data into the graph
• Determine the minimum viable set of data to bootstrap the graph
• Retain federation of data internally
• Data authorities keep editorial and publishing control as before
• If we can prove out a knowledge graph of federated data internally, we can usesame approach to link to customers data and open data
![Page 21: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/21.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 21/27
, -,
THOMSON REUTERS KNOWLEDGE GRAPH STAT
• Knowledge Graph: 2.35B triples- Metadata, Organisations, People: 2.27B triples- Inferred Data, generated with SPIN rules [reverse predicates etc.]: 78.3M triples
• Compared to other large open data sets:- Wikidata: 367M triples
- DBPedia: 474M triples- Freebase: 2B triples- UniProt: 17B triples
![Page 22: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/22.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 22/27
, -,
KNOWLEDGE GRAPHS PROVIDEMANY VIEWS TO ANSWER MANY QUESTIONS
• Gives us the ability to provide many lenses over the graph
• Query for absolute facts• Patents issued by a company, litigation history, market capitalisation history
• Also make inferred and abstract connections• Sort by litigation history within an industry sector weighted by market capitalisation
• Combine absolute facts with inferred/abstract connections
![Page 23: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/23.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 23/27
, -,
KNOWLEDGE GRAPHS PROVIDEMANY VIEWS TO ANSWER MANY QUESTIONS
• Iterate and build layers of queries of increasing sophisticated/complexity to innew facts
• Handle relative truth of facts and data - according to their source• Ability to utilise the facts relevant to your product or question• Adding additional perspectives as relevant
![Page 24: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/24.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 24/27
, -,
KNOWLEDGE GRAPH USE CASE
![Page 25: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/25.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 25/27
![Page 26: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/26.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 26/27
![Page 27: Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015](https://reader030.vdocuments.mx/reader030/viewer/2022021103/577ca7ab1a28abea748c81e7/html5/thumbnails/27.jpg)
8/20/2019 Creating the Thomson Reuters knowledge graph and open permID - ODI Summit 2015
http://slidepdf.com/reader/full/creating-the-thomson-reuters-knowledge-graph-and-open-permid-odi-summit-2015 27/27