At Center Stage IV: Ontotext Webinars About How GraphDB Levels the Field Between RDF and Property Graphs

A series of blog posts focusing on major Ontotext’s webinars and how they fit into the bigger picture of what we do. In this post, we talk about how GraphDB has eliminated the main limitations of RDF vs LPG by enabling edge properties with RDF-star and key graph analytics within SPARQL queries with the Graph Path Search plug-in.

November 5, 2021 8 mins. read Gergana PetkovaAtanas KiryakovAtanas Kiryakov

This post continues our series in which we want to provide an overview of what we do and how our webinars fit into it. In the previous post, we talked about data virtualization and how Ontotext’s RDF database for knowledge graphs GraphDB provides the tools for the full journey from graphs to relational tables and back. All of our webinars are available on demand, so if we pique your interest with any of these posts, you can request a free recording.

Here, we want to draw your attention to two webinars that focus on how GraphDB levels the field between RDF and LPG stacks with the help of RDF-star support and the Graph Path Search plug-in. Given all the advantages of the RDF model regarding enterprise data management, one no longer has good arguments why to bother with LPG at all.

Read on and watch the webinars to get convinced.

The Background Story

As we’ve said again and again, we believe that knowledge graphs are the next generation tool for helping businesses make critical decisions, based on harmonized knowledge models and data derived from siloed source systems. Due to the huge value generated by their data standardization and semantic modeling capabilities, knowledge graphs are most often associated with data integration, linking, unification and information reuse. But these tasks are only part of the story. As more and more organizations are turning to knowledge graphs for better enterprise knowledge management, data and content analytics, search and graph exploration become key tools for successfully utilizing knowledge graphs.

In this age of graphs, there are two main graph data models, each of them with distinct characteristics:

  • RDF: W3C’s Resource Description Framework and related standards for schema, query  and validation languages like RDFS, OWL, SPARQL and SHACL;
  • LPG: the labeled property graphs, rooted in Apache’s ThinkerPop “graph computing framework” and GREMLIN “graph traversal machine and language”.

These models originate from different use cases: distributed knowledge representation and open data publishing on the web vs graph analytics designed to be as easy to start with as possible. There have been many comparisons of the strengths and limitations of knowledge graphs vs property graphs (this is our take on it).

Long story short, RDF is recognized as the better option (if not the only one) for knowledge graphs, because its web-native syntax supports data exchange and sharing and because its formal semantics allows for easy alignment of meaning and structure across sources, unified views and unambiguous interpretation.

On the other hand, LPGs lack many features that are an absolute must for enterprise data management, e.g., standard schema language. Still, for many years, two main advantages of property graphs have been pointed out: they can deal with properties on edges in the graph and they are good for graph traversal.

In two of its 9.x releases, GraphDB has eradicated these two limitations by upgrading its modeling expressivity with RDF-star/SPARQL-star and by enabling key graph analytics within SPARQL queries with the Graph Path Search plug-in. Both these new features have already been adopted in partner tools:

Now, let’s dive in and look into each of these webinars.

GraphDB 9.2 Supports RDF-Star to Match the Expressivity of Property Graphs

In our first webinar GraphDB 9.2 Supports RDF-Star to Match the Expressivity of Property Graphs,our CTO Vassil Momtchev demonstrates the various options of modeling complex relationships that have multiple attributes with plain RDF, e.g., reification and singleton properties. He walks you through what you need to know about RDF, the different modeling challenges we’ve seen over the years and how to address them. RDF-start overcomes all these challenges. It also goes beyond the expressivity of property graphs, where you can attach key-value pairs to relationships and allows a more efficient representation of scores, weights, temporal restrictions and provenance information.

You will learn more about statement level metadata, the pros and cons of RDF-star, how SPARQ-star works and how different RDF engines implement RDF-star. Vassil introduces the new GraphDB features related to RDF-star/SPARQ-star and goes over some practical examples and performance results. He shows how RDF-star brings the simplicity and usability of property graphs without sacrificing the essential semantics that enables correct interpretation and efficient management of the data.

Interesting attendee question: Should I model my data, such as start and end date, as metadata with embedded triples or as N-ary concepts?

Ontotext Answer: If we take again the example of Abraham Lincoln being the President of the United States (see the visual above), our recommendation is to model the President of the United States in the ontology model in order to enable inference and other interesting things. This is a N-ary relationship concept that would be very clear for every user of the model – that “president” is a type of position with a “start date” and “end date”. And we would recommend to model other metadata (such as the reference and where this information comes from) as RDF-star embedded triples because maybe this is something that doesn’t need to be exposed for every user but only to the people who are interested in the statement.

New call-to-action

 

Graph Path Search with GraphDB 9.9 and metaphactory 4.3

The second webinar in focus is Graph Path Search with GraphDB 9.9 and metaphactory 4.3. It demonstrates how Ontotext and its technology partner metaphacts can generate great value on top of knowledge graphs in analytical use cases through graph path search and interactive visualization.

In the first part of the webinar, our speaker from Ontotext, Tomas Kovachev, explains why graph path search is a computationally expensive task and presents our graph path search implementation. He also provides a comparison of how different RDF and property graph databases implement it and how GraphDB extends the SPARQL 1.1 standard to fully support all significant graph path search use cases. On top of that, he tests GraphDB and other engines against LDBC’s Semantic Network Benchmark to give you a feeling for the scalability and the efficiency of our implementation.

In the second part, our guest speaker Sebastian Schmidt, CEO at metaphacts (the first one to adopt our new implementation), demonstrates how the graph path search algorithm available in GraphDB 9.9 is exposed to end users with the metaphactory 4.3 release. He shares a specific use case taken from the clinical trial domain: How a researcher can find connections between study investigators and various other resources such as targets and diseases, which will allow them to further investigate discovered studies or medications

You will learn more about graph path search fundamentals, the features in the “vanilla” SPARQL 1.1 limiting the exploration of paths in a graph and how the new extension in GraphDB 9.9 enables complex graph pattern search in full compliance with the SPARQL specification. You will also see a demonstration of how graph path search is exposed to non-techie end users in the metaphactory 4.3 release.

Interesting attendee question: How does this search for the shortest path work when there is a cyclic relation in the graph? Do you try breaking the cycle first? 

Ontotext answer: Graphs can represent limitless relationships between resources, therefore cycles have to be taken into account in order to avoid endless graph traversals. Our path finding algorithm is based on the level-by-level adjacent node traversal of Breadth First Search but, unlike the original algorithm, it keeps track of all the visited nodes (at all depths) and the relations leading to them. This allows us to not only avoid cyclical relationships but also to find whether a given resource can be reached via other resources and relationships.

New call-to-action

 

Wrap up

That covers the two webinars that we wanted to present to you today.

Take home summary:

  • RDF-star provides an easy way to model and query edge properties;
  • RDF-star allows for much more expressive metadata on relationships than LPGs;
  • GraphDB’s implementation of RDF-star is very space efficient and performant;
  • Finally, there is a comprehensive extension of SPARQL for graph-path search, which is fully compliant with the standard and addresses all tasks, e.g., all path search;
  • GraphDB’s graph-path search extension proves its efficiency on the LDBC Social Network Benchmark.

In a nutshell, given all the advantages of RDF, and GraphDB in particular, regarding knowledge graph applications, one has no longer any reasons to bother with property graphs, which are not fit for efficient enterprise data management. Managing your data and analyzing it can go hand in hand. You can use the same knowledge graph platform that lets you put your data together in a sustainable way for analytics purposes as well.

We hope you’ve heard enough to want to dive straight into one of these webinars or why not both?

Watch out for our next post in this series At Center Stage: Ontotext’s Webinars or visit our Webinars directly to check out what is upcoming!

Article's content

Content Manager at Ontotext

Gergana Petkova is a philologist and has more than 15 years of experience at Ontotext, working on technical documentation, Gold Standard corpus curation and preparing content about Semantic Technology and Ontotext's offerings.

Atanas Kiryakov

Atanas Kiryakov

CEO at Ontotext

Atanas is a leading expert in semantic databases, author of multiple signature industry publications, including chapters from the widely acclaimed Handbook of Semantic Web Technologies.

Knowledge Graphs: Redefining Data Management for the Modern Enterprise

Read this post about some of the primary problems of today’s enterprise data management and how knowledge graphs can solve them

Knowledge Graphs: Breaking the Ice

Read about the nature and key characteristics of knowledge graphs. It also outlines the benefits of formal semantics and how modeling graphs in RDF can help us easily identify, disambiguate and interconnect information

GraphDB in Action: Navigating Knowledge About Living Spaces, Cyber-physical Environments and Skies 

Read about three inspiring GraphDB-powered use cases of connecting data in a meaningful way to enable smart buildings, interoperable design engineering and ontology-based air-traffic control

Your Knowledge Graph Journey In Three Simple Steps

A bird’s eye view on where to start in building a knowledge graph solution to help your business excel in a data-driven market

Data Management Made Easy: The Power of Data Fabrics and Knowledge Graphs

Read about the significance of data fabrics and knowledge graphs in modern data management to address the issue of complex, diverse and large-scale data ecosystems

GraphDB in Action: Powering State-of-the-Art Research

Read about how academia research projects use GraphDB to power innovative solutions to challenges in the fields of Accounting, Healthcare and Cultural Heritage

At Center Stage VIII: Ontotext and Enterprise Knowledge on the Role of Knowledge Graphs in Knowledge Management

Read about our partnership with Enterprise Knowledge and knowledge management as an essential business function and lessons learned from developing content recommenders using taxonomies and GraphDB.

At Center Stage VII: Ontotext and metaphacts on Creating Data Fabrics Built on FAIR Data

Read about our partnership with metaphacts and how one can use the metaphactory knowledge graph platform on top of GraphDB to gain value from their knowledge graph and accelerate their R&D.

At Center Stage VI: Ontotext and Semantic Web Company on Creating and Scaling Big Enterprise Knowledge Graphs

Read about our partnership with Semantic Web Company and how our technologies complement each other and bring even greater momentum to knowledge graph management.

At Center Stage V: Embedding Graphs in Enterprise Architectures via GraphQL, Federation and Kafka

Read about the mechanisms for building a big enterprise software architectures by embedding graphs via GraphQL, Federation and Kafka

Ontotext’s Perspective on an Energy Knowledge Graph

Read about how semantic technology can advance energy data exchange standards and what happens when some energy data is integrated in a knowledge graph.

At Center Stage IV: Ontotext Webinars About How GraphDB Levels the Field Between RDF and Property Graphs

Read about how GraphDB eliminates the main limitations of RDF vs LPG by enabling edge properties with RDF-star and key graph analytics within SPARQL queries with the Graph Path Search plug-in.

At Center Stage III: Ontotext Webinars About GraphDB’s Data Virtualization Journey from Graphs to Tables and Back

Read this second post in our new series of blog posts focusing on major Ontotext webinars and how they fit in the bigger picture of what we do

At Center Stage II: Ontotext Webinars About Reasoning with Big Knowledge Graphs and the Power of Cognitive Graph Analytics

Read this second post in our new series of blog posts focusing on major Ontotext webinars and how they fit in the bigger picture of what we do

At Center Stage I: Ontotext Webinars About Knowledge Graphs and Their Application in Data Management

Read the first post in our new series of blog posts focusing on major Ontotext’s webinars and how they fit in the bigger picture of what we do

The Gold Standard – The Key to Information Extraction and Data Quality Control

Read about how a human curated body of data is used in AI to train algorithms for search, extraction and classification, and to measure their accuracy

Study of the Holocaust: A Way Out of Data Confusion

Learn how a ML algorithm trained to replicate human decisions helped the EU-supported EHRI project on Holocaust research with record linking.