Delivering Linked Data Training to Data Science Practitioners
Post on 27-Jan-2015
102 Views
Preview:
DESCRIPTION
Transcript
Online Learning and Linked Data – Lessons Learned and Best Practices (WWW’2014)
Delivering Linked Data Training to Data Science Practitioners
Marin Dimitrov (CTO)
Contents
• Ontotext’s experience with Linked Data trainings
• Lessons learned
• The EUCLID social media monitoring platform
#2 Online Learning and Linked Data – Lessons Learned and Best Practices WWW’2014
Apr 2014
About Ontotext
• Provides products & solutions for semantic metadata management
– Multi-lingual text analytics
– Semantic search
– Linked Data management
• Major clients and industries
– Media & Publishing
– Health Care & Life Sciences
– Cultural Heritage
– Government
– Recruitment
#3 Online Learning and Linked Data – Lessons Learned and Best Practices
WWW’2014
Apr 2014
The Need for Linked Data Training
• We realised that “educating the market” was a required prerequisite for selling the technology
– 2009: started with product specific training modules
– 2011: training also covers general Linked Data / Semantic Web topics
• Current training curriculum
– Product specific training
– Introduction to Semantic Web
– RDF/S, OWL, ontologies, inference
– SPARQL & RDF databases
– Linked Data & applications
#4 Online Learning and Linked Data – Lessons Learned and Best Practices WWW’2014
Apr 2014
Linked Data Training for Data Science Practitioners
• Big organisations / enterprises
– BBC
– British Museum
– UK parliament
– One of the world’s leading financial news organisations
– One of the leading monthly magazines focussed on business & finance
– A global academic publishing company
– One of the biggest & most visited museums in the US
#5 Online Learning and Linked Data – Lessons Learned and Best Practices WWW’2014
Apr 2014
Linked Data Training for Data Science Practitioners
• Conferences & events
– NYC, San Francisco, London, Stavanger
• On-premise & online training
#6 Online Learning and Linked Data – Lessons Learned and Best Practices WWW’2014
Apr 2014
Lessons Learned
• #1 Be prepared for a mixed audience
– Often a mix between managers, developers & DBAs
– A mix between groups with different prior expertise
– Not easy to tailor the message to the audience needs
• #2 Speak the language of the audience
– Comparison with relational databases / SQL
– Linked Data applications & APIs for developers
– Success stories for managers
#7 Online Learning and Linked Data – Lessons Learned and Best Practices WWW’2014
Apr 2014
Lessons Learned
• #3 Strike the right balance between too much & too little theory
– Proper theoretical foundations are important
– … but difficult to plan in advance
– Adjust the training content on the fly
• #4 Success stories & reference architectures are important
– Focus on how and why Linked Data is used in various domains, not only on what it is
– Understanding of the business domain is important
– Organisations want to replicate BBC’s initial with Linked Data applied in production
#8 Online Learning and Linked Data – Lessons Learned and Best Practices WWW’2014
Apr 2014
Lessons Learned
• #5 Engage the customer pre- and post-training
– Pre-training: Helps to understand better the audience of the training
– Pre-training: customers can prepare better - background reading, questions, etc.
– Post-training: clarifications, Q&A sessions, further reading
• #6 Record the trainings / webinars
• #7 Bottom-up approach is needed sometimes
– Q&A sessions / open discussion, instead of pre-planned course
– More of a consulting session than a standard training
#9 Online Learning and Linked Data – Lessons Learned and Best Practices WWW’2014
Apr 2014
EUCLID SOCIAL MEDIA MONITORING PLATFORM
#10 Online Learning and Linked Data – Lessons Learned and Best Practices WWW’2014
Apr 2014
EUCLID Social Media Monitoring Platform
• Goals
– Identify “hot topics” related to Linked Data & Semantic Web
– Provide topics for extending & refining the existing EUCLID Linked Data curriculum
• Approach
– Monitor Twitter, selected W3C mail lists, LinkedIn discussion groups, SlideShare uploads
– Integrate & RDF-ize the data coming from various channels
– Automatically extract topics & named entities (text analytics)
– Visual analytics on top of the integrated data
#11 Online Learning and Linked Data – Lessons Learned and Best Practices
WWW’2014
Apr 2014
EUCLID Social Media Monitoring Platform
• Knowledge base
– FreeBase + GeoNames
– Semantic Web conferences metadata
– Dublin Core, FOAF, SCIOC, SKOS topics
– Manually refined set of Top200 influencers in the Linked Data / Semantic Web community
• Text mining, entity resolution & interlinking
• Data visualisation
#12 Online Learning and Linked Data – Lessons Learned and Best Practices WWW’2014
Apr 2014
EUCLID Social Media Monitoring Platform
#13 Online Learning and Linked Data – Lessons Learned and Best Practices WWW’2014
Apr 2014
EUCLID Social Media Monitoring Platform
#14 Online Learning and Linked Data – Lessons Learned and Best Practices WWW’2014
Apr 2014
THANK YOU!
#15 Online Learning and Linked Data – Lessons Learned and Best Practices WWW’2014
Apr 2014
top related