Make your data accessible useful trusted
Knowledge graphs that put you in control and deliver value in weeks.
Knowledge graphs from Eigen
Meet Murray, Eigen’s CEO
- Mapping data from different, disparate systems.
- Digitising information hidden in documents.
- Tying up valuable engineering time hunting for data.
Murray has been leading the charge on ‘taming the data monster’ with knowledge graph technology rather than trying to boil the ocean with a data lake.
Fifteen years working at the coal face with offshore oil and gas data and systems has shown us how operators need more dynamic tools than databases and data lakes
A knowledge graph is a network of nodes of data – and pointers to the data. The pointers explain the relationships with the data, so it works a bit like the brain, connecting data with context for when it’s needed.
We build Knowledge Graphs for oil and gas to help operators maximise value from their data, by:
Saving time hunting for data
Breaking down data silos
Increasing confidence in data
Driving efficiencies through automated workflows and reporting
Making more data available and accessible
Data is one of your most precious assets - and one of our biggest challenges to solve
For many operators, their data has become unwieldy and complex, having evolved over years and even decades. It’s dynamic, always changing – and increasing as more equipment is instrumented.
Are you struggling with any of these common problems?
1
Data in silos across multiple systems, databases, organisations.
2
Legacy data that few understand.
3
Different naming conventions.
4
Information about assets dispersed across many systems.
5
Data barriers to automating analytics workflows across all assets.
At Eigen we build knowledge graphs that work like the human brain to connect up data across multiple sources. We help to mask the complexity of nomenclature and location to present data the way engineers and operators need it – in a report, as a calculation input, feeding an automated workflow.
And we build knowledge graphs fast, use case by use case to unlock value fast.
Industrial knowledge graphs that create order and value
Knowledge graph technology offers a faster and more scalable approach to managing data and maximising value.
Eigen brings 15 years of domain expertise in O&G systems and graph building experience to help
operators get more value from their data faster.
What we have learnt is that designing a knowledge graph for the entire organisation is just as unlikely to succeed as designing and building a large data lake. Building a knowledge graph works best use case by use, building incrementally and delivering value incrementally.
Five benefits of Eigen knowledge graphs
1
They point to your source data, no need to move or copy any data
2
They bring context with the data to explain the relationships
3
They scale as more knowledge is used to augment the graph
4
There aren’t predefined boundaries
5
They don’t lock you in to a specific platform or vendor – you retain control of the knowledge
Your questions about knowledge graphs
These are the top questions we get asked about knowledge graphs. If you have others, try our FAQs.
Your IT teams may have other questions – about the architecture, digital security and data governance.
How is a knowledge graph different from a data lake?
A data lake is a cloud-based data repository designed to ingest and contextualize data from source systems, effectively creating a copy of the original data. Eigen knowledge graphs do not ingest data, instead they provide pointers or references to the source data within the systems of record.
How long does a knowledge graph take to build?
Our approach to creating the graph is incremental and organic. Because there is no predefined schema, the smallest possible graph is created to support the first use case, and then it incrementally grows to support subsequent use cases. By following this methodology, the creation of the knowledge graph becomes just another task in the implementation of a value-adding use case.
Does a knowledge graph become an uncontrollable monster?
Knowledge Graphs need governance and robust processes and is the ultimate responsibility of the graph owner. Information standards such as ISO15926, IEC81436, NORSOK and CFIHOS can be designed into the graph. Eigen can provide tools to help enforce standards as the graph evolves over time. Any unwanted changes can be rolled back to preserve the integrity of the graph.
Simple for you to build, quick to payback
Start building in minutes
As engineers, we are natural problem solvers; we like to get our hands dirty – and with knowledge graph technology we can do just that – and so can you.
Knowledge graphs are easy to start building because there is no need to predefine boundaries or structures as with a database or data lake. Neither are you locked into a platform or someone else’s architecture.
See how easy it is to build a knowledge graph that creates value in under three minutes.
How to build a knowledge graph
90 days to value – or less
Building a knowledge graph, as part of the development of a value-adding use case, works best using the agile method with short sprint cycles and iterative reviews.
Our agile approach is to focus on a single pain point and build out a single use case for the knowledge graph, identifying source data and the relevant calculations and functions.
The key difference with knowledge graph projects lies in the ability to quickly demonstrate small chunks of value, use case by use – rather than waiting months for a new platform and migration.
Eigen’s agile approach
There are three iterative phases for any Eigen project: Define objectives, Design concept and Run pilot.
1
Define objective
• Set clear business goals
• Focus on problem or pain point
• Agree metrics
2
Define concept
• Identify data sources
• Consult stakeholders
• Investigate architecture and security requirements
• Design system
3
Run pilot
• Build solution*
• Test and evaluate
• Plan next steps, including rollout or scaleup
*Eigen can provide trial licenses during pilot phase.
Knowledge graphs that deliver
We have built knowledge graphs for safety use cases, like barrier health monitoring, plant shutdown analysis, and for automated reporting.
Their ability to pull in hundreds of thousands of data points from multiple different sources – with context – in seconds lends itself to numerous applications.
Very often we find working with a client that we can address part of a use case with what’s already been built into their knowledge graph – meaning we can deliver value much earlier.
eigen case study
Automated Blowdown Verification
“It’s saved us up to a week of an engineer’s time and already proved useful in trouble shooting and risk evaluation that we didn’t think of in advance. It’s given us much more than we aimed for.”
Gauthe Aaboen, Senior Engineer Operations, Akerbp
FAQs
You may have questions about knowledge graph technology. We have tried to answer some here – and have more FAQs for IT teams here.
If you have other questions, email Murray Callander or book a demo.
Do you have to create a graph to represent my entire operation before it becomes useful?
No. The knowledge graph becomes useful and delivers value if it can support the implementation of the initial use case, therefore only the objects and relationships required to support such use case are
required to be represented in the graph.
Are you copying or ingesting the data?
Neither. Eigen knowledge graphs do not ingest any data, which remains in the existing systems of record (the “functional system”). The knowledge graph captures key attributes and metadata related to each object and the relationships between objects. The Knowledge Graph is not a data lake and does not replace the functional system
What is a pointer?
A pointer or reference is a piece of metadata, usually a url, stored within a knowledge graph node that provides a link to the asset data contained within the systems of record associated with such node.
There can be pointers to workorders, real time data tags, documents or asset registers. It can also be paramaters that, when submitted to an API, return the relevant data from the source system.
How does the knowledge graph scale?
We build our Knowledge Graphs in Neo4j. This is a mature technology that is proven to scale indefinitely (https://neo4j.com/product/neo4j-graph-database/scalability). However, when scaling the graph, technology is not really the challenge – it’s the graph structure that determines if it will still perform well when it contains a hundred, a million or a billion nodes. At Eigen we understand how graph databases work, how to optimise queries and, most importantly, we understand the industrial domain and how to structure the data.
Won’t the knowledge graph be slow?
Eigen’s knowledge graph is powered by Neo4j graph technology, which exhibits excellent elasticity and scalability. Additionally, the configuration of our knowledge graph, includes generic elements designed to optimize the efficiency of queries. Specific use cases that require online access to source datasets could impact the overall performance of the EAP, but in these cases, caches can be set up to speed up the response time.
Does the knowledge graph support multiple hierarchies?
Absolutely yes. Unlike a tree structure which can only support one hierarchical organization, the knowledge graph is not limited by the number of relationships that each object can have and thus can model multiple concurrent hierarchies.
Can Eigen’s knowledge graph use our own equipment naming convention?
Yes, the client’s preferred nomenclature or asset naming convention can be used within the knowledge graph. Objects created using this nomenclature generally co-exist with other abstract objects required to support specific use cases.
Once we start building a knowledge graph, are we dependent on Eigen?
No. The knowledge graph is your property and is configured on Neo4j open-source technology accessible using Cypher. The user is free to query the graph using external tools therefore is not dependent on Eigen. We actively encourage our customers to train up experts within their business who can become the custodians of the graph and the in-house experts in how to use it.
Got another question? Email Murray Callander.