The challenge

Combining complex connected data sets

Our world is deeply connected, but these connections are not always obvious. When you imagine your connections, you may think of friends, family and places. In business, this may be competitors, suppliers, where you operate, who you sell to and who you want to sell to. Often these data points come from multiple sources which on their own, only tell part of the story. But, being able to connect all of these data points together and articulate the relationships between them gives a richness to the story that enables us to see the whole picture, and in doing so make better data-driven decisions.

StellarGraph have developed cutting-edge tools to reveal actionable insights from data and expose its real value. We are engineers, data scientists, researchers, product managers and UX designers, all driven to build effective, purposeful technology that solves real-world problems.

Thomas H Davenport said that "every company has big data in its future and every company will eventually be in the data business". The challenge though, is leveraging the power of this data to drive insights that really matter and help us make high-risk, high-impact decisions with confidence. This is especially difficult when we are not only dealing with vast quantities of data but looking to combine complex connected data sets from multiple sources, and in varying formats. The useful insights are often hidden from the human eye, and more traditional two-dimensional graph representation that capture data in table format just don't cut it. The problem spans industries from financial services looking to detect fraud and money laundering, shipping and transport seeking to contain illegal trafficking, to law enforcement risk and threat protection.

The solution can be found in graph analytics.

Our response

Capturing data as a graph

Capturing data as a graph enables the context and rich, relationship-driven structure or multiple data sources to be modelled so that we can see the full picture. StellarGraph have developed key technologies to answer the challenges in assessing and understanding insights from connected data.

Many real-world datasets can be naturally represented as networks or graphs, with nodes representing entities and links representing relationships or interactions between them.

Fig.1 Many real-world datasets can be naturally represented as networks or graphs, with nodes representing entities and links representing relationships or interactions between them.

Fig. 2 When there are hundreds of entities, or thousands, or millions of entities, things can escalate quickly and soon enough; you have one big hairball.

The StellarGraph Library is a commercial grade, open-source graph machine learning library written in Python. We developed the library to democratise machine learning on networks and graphs for data scientists, developers and researchers in government or industry wanting to experiment with graph machine learning techniques and/or apply them to their network-structured data.

The StellarGraph Platform is a commercial grade, distributed platform that allows us to ingest data, create graphs and apply performant machine learning at scale to billions of data points. This means we can leverage machine learning where it counts - on production-scale data, to get greater value insight from data assets.

The StellarGraph Visualisation tool is a cloud application for exploring huge networks, extracting insights and sharing knowledge to enable data-driven decisions. The software is designed to help analysts understand massive network datasets in a visual way, empowering users to understand their data and share discoveries with others.

The results

Machine learning processing connected data at scale

There are countless examples of rich, network data sets being under-utilised across government and industry. StellarGraph technologies deliver the ability to automate or semi-automate manual processes to improve efficiency and enhance variable output.

Machine learning techniques enable us to process the connected data at scale to get greater clarity on the contents, reveal hidden connections, prioritise what is important and find with greater speed what we are looking for. What's more, StellarGraph enables people to have trust in machine learning outcomes, because our tools infiltrate the data black box to cause interpretability. The impact is that teams of people can reach agreement quickly on which actions to take and can execute these actions with confidence. This puts StellarGraph in a position to support government and industry in solving real problems.

StellarGraph's focus is also on releasing cutting-edge research in an applied, user-centric way. We've done the hard work, resulting in tools that empower data scientists to build, test and experiment with machine learning models to create value from network data. Simply, StellarGraph is contributing to strengthening Australia's data science capability.

Version 1.0 of the StellarGraph open source Python Library , was released in May, delivering optimised capability and speed as well as a suite of powerful new algorithms and demos for machine learning on graphs.

Do business with us to help your organisation thrive

We partner with small and large companies, government and industry in Australia and around the world.

Contact us now to start doing business

Contact Data61

Your contact details

First name must be filled in

We'll need to know what you want to contact us about so we can give you an answer.