Towards a productive Linked Data environment within Enterprises
1 like226 views
The document discusses enabling a productive Linked Data environment within enterprises. It outlines three common unwanted situations: seeing Linked Data as only strategic and not tactical, viewing it as just another ETL process, and focusing too much on formalisms. The document then proposes five success factors: using domain-specific metaphors, joining developer communities, adopting agile practices, taking a community-focused approach, and being pragmatic. The conclusion restates the unwanted situations and success factors as takeaways for productive Linked Data projects within companies.
Towards a productive Linked Data environment within Enterprises
- 1. Towards a productive Linked Data environment within Enterprises Andreas Both 2019-05-22, Leipziger Semantic Web Tag (LSWT 2019)
- 2. LSWT 2019: Towards a productive Linked Data environment within Enterprises Andreas Both Motivation for the talk • share experience and insights • provide options of actions • help to overcome common misconceptions Disclaimer Imagesource:pixabay.com/illustrations/signs-right-wrong-good-bad-1172209/–License:PixabayLicense. 2 of 31
- 3. LSWT 2019: Towards a productive Linked Data environment within Enterprises Andreas Both Motivation for the talk A talk about “Linked Data towards action” ... isn’t this like 5 years ago? Positive • improved tools/toolchains • more open data sets • many new vocabularies • in general: increased understanding of Linked Data engineering processes Negative • many companies still struggle on taking advantage of linked enterprise data • applying the Linked Data paradigm still no common approach 3 of 31
- 4. LSWT 2019: Towards a productive Linked Data environment within Enterprises Andreas Both Spotlights 4 of 31
- 9. LSWT 2019: Towards a productive Linked Data environment within Enterprises Andreas Both Recapitulation Main goal Web of (linked) documents → Web of Linked Data ⇒ make the (data) world a better place! Core Ideas of Linked Data 1. identify things using URIs/IRIs 2. use URIs/IRIs to refer to data 3. provide semantics for data, make semantics accessible 4. interlink data stores → break established data-access approach in enterprises ⇒ a huge change in understanding and behaving is expected from teams/organisations! 9 of 31
- 10. LSWT 2019: Towards a productive Linked Data environment within Enterprises Andreas Both About change management, it is. While changing the information system paradigm ... • change of project goals and metrics • change of project execution • change of required skills • change of technology stack • change the common understanding of “finished” • change of pattern (team communication, software design, iteration slicing, . . . ) • change the power balance (long-term vs. short-term, visual vs. backend, . . . ) • . . . → change management is a very important issue Image“Yoda”byPaulHudsonislicensedunderCCBY2.0 10 of 31
- 11. LSWT 2019: Towards a productive Linked Data environment within Enterprises Andreas Both Change Management Incomplete questions to initialize and drive a change process: • What might cause misconceptions? • What will make the change stick? • What might block the change? Observation Mostly about changing the behavior of individuals! 11 of 31
- 12. LSWT 2019: Towards a productive Linked Data environment within Enterprises Andreas Both Experienced Situations 12 of 31
- 13. LSWT 2019: Towards a productive Linked Data environment within Enterprises Andreas Both Experienced Situations Feedback: “Linked Data technologies solve a strategic problem” • is enabler for additional value chain • requires changing of paradigms with long-term effects Observations and Consequences hard for companies to invest enough into strategic approaches 13 of 31
- 15. LSWT 2019: Towards a productive Linked Data environment within Enterprises Andreas Both Experienced Situations Feedback: “Linked Data is yet another ETL data integration process” • data needs to be transformed • for good experience a common approach is to copy data into a triplestore Observations and Consequences classification of Linked Data projects similar to ETL projects (well-know in enterprises for 20+ years) 15 of 31
- 17. LSWT 2019: Towards a productive Linked Data environment within Enterprises Andreas Both Experienced Situations Feedback: “The whole movement seems to follow a formal approach.” • ontologies provide sound theoretical foundations • during data modelling we require precise statements 17 of 31
- 19. LSWT 2019: Towards a productive Linked Data environment within Enterprises Andreas Both Success Factors 19 of 31
- 20. LSWT 2019: Towards a productive Linked Data environment within Enterprises Andreas Both Success Factors Find matching metaphors • less: reuse of common linked data terminology • more: domain specific, precisely matching naming → positive impact: easier for customers to adopt your thinking and better expectation management Example: “data silos” 20 of 31
- 22. Imagesource:pixabay.com/photos/silo-grain-texas-southwest-tower-1476422/–License:PixabayLicense.
- 25. LSWT 2019: Towards a productive Linked Data environment within Enterprises Andreas Both Success Factors Join a community of developers (product development) • less: disconnection from product development • more: advantages of JSON-LD, openAPI extension, . . . → positive impact: bottom-up movement starting Derived research questions improve integration into software engineering process 25 of 31
- 26. LSWT 2019: Towards a productive Linked Data environment within Enterprises Andreas Both Success Factors Agile Approach • less: long-term planning, high ramp-up costs, top-down approach • more: agility and iterations → increased transparency and faster value transfer Derived research questions tools for change management of ontologies 26 of 31
- 27. LSWT 2019: Towards a productive Linked Data environment within Enterprises Andreas Both Success Factors Community Ontology Management Approach • less: centralized approach, gate-keeper • more: community approach, community management → increased commitment in development teams, increased quality Derived research questions improve tooling for community management, co-evolution, self-services, . . . Image“bouncers2”byCharlesLeBlancislicensedunderCCBY-SA2.0 27 of 31
- 28. LSWT 2019: Towards a productive Linked Data environment within Enterprises Andreas Both Success Factors Pragmatic Approach • less: theoretical, formal, meta-level discussions • more: pareto-optimal actions → positive impact: first impact quickly achieved Derived research questions improve method and tool capabilities w.r.t. uncertainty 28 of 31
- 29. LSWT 2019: Towards a productive Linked Data environment within Enterprises Andreas Both Conclusions and Take Away 29 of 31
- 30. LSWT 2019: Towards a productive Linked Data environment within Enterprises Andreas Both Conclusions • 3 unwanted situation ◦ w.r.t. strategic importance ◦ w.r.t. ETL processes ◦ w.r.t. formalisms • 5 action items which might be success factors ◦ metaphors ◦ community of developers (community of practice) ◦ adopt agile methodology ◦ community focus ◦ pragmatic approach 30 of 31
- 31. LSWT 2019: Towards a productive Linked Data environment within Enterprises Andreas Both Take Away • take on classification of challenges you might be confronted while working in the field of Linked Data • insights into some common challenges during the execution of Linked Data projects • possible action items during the execution of a Linked Data driven project → towards a productive Linked Data environment within enterprises Prof. Dr. Andreas Both andreas.both@hs-anhalt.de Anhalt University of Applied Sciences linkedin.com/in/andreas-both-9426722 31 of 31