This content is locked

This content is no longer available.

Video content locked

March 12, 2021, 08:00 PM

Modelling Physical Systems with Graph Databases

A talk by Mike Morley and Peter Tunkis
Arcurve & Arcurve, Inc.

Speaker expertise

Focused on applying technology to industry and enterprise ... Creating and empowering teams through a highly collaborative team first leadership approach, driven by a deep fascination of non-hierarchical organizational design patterns. Works with a wide spectrum and diverse of technologies and platforms... javascript/nodeJS, c#, SQL, python, neo4j/cypher, databricks, messaging patterns, containerized architectures, kimball datawarehouse/olap patterns, nosql databases... B.A.S.c. Geological Engineering University of Waterloo...

I have a long career in working with spatial data. This covers the range of standard GIS (ESRI and Geoserver, 3d modelling for mining operations, complex geological modelling, buildings and structures, finite element and numerical analysis.

So it was incredibly exciting when Neo4j first introduced spatial capabilities. While these are currently limited compared to what something like ESRI offers in terms of standard GIS analysis such as buffering etc., the fact that neo4j supports both 3d spatial and 3d cartesian coordinates opens up a whole realm of possibilities that will ultimately transcend what a typical GIS can do.

This is because neo4j due to its graph model of nodes and relationships when combined with being able to assign 3d spatial and Cartesian coordinates to nodes becomes both a GIS and what runs behind most CAD and numerical models. Standard CAD and GIS systems use tabular data structures that are converted to graph structures in memory to do their work. Generally the data aspect and the structural/spatial/3d aspects are stored in separate table structures. Autocad and Revit actually use completely separate data and structural data stores to model buildings and other complex 3d objects.

Neo4j with its spatial capabilities now makes it possible to actually fully combine the 3d structure, spatial structure and data into a single model. It also being a graph database then has all the power that affords, plus the text analytics, Lucene full text engine and AI/ML libraries. This suddenly means that it becomes possible to model spatial surface models such as Open Street Maps, but the buildings on those maps, the boreholes and subsurface data that make up geological or geotechnical data, and all of the reports, design documents and other structured and unstructured data that are associated with any complex engineering structure.

This has the potential to be incredibly powerful and very disruptive to the world of engineering and geology.

Pete Tunkis a member of the Arcurve Data Science team will be joining me to help illustrate some of the potential we see...

Join the discussion

You need to be registered as an attendee in order to comment on this talk.


Categories covered

Proudly supported by

Want to sponsor this event? contact us.