An introduction to Raphtory, the temporal graph engine
Naomi Arnold is a postdoctoral research assistant at the Network Science Institute of Northeastern University in London and Ben Steer is the CTO and Co-Founder at Pometry as well as an honorary researcher at Queen Mary University of London
Data is at the heart of decision-making today, and graphs are firmly embedded in the modern data stack. From fraud detection and drug discovery to market and supply modelling, graphs enable previously unachievable insights. However, while graph analytics platforms are increasingly used, most applications and solutions overlook a crucial element: time. Raphtory is an analytics platform which combines the capabilities of a graph structure with time-series analysis, providing a powerful model for extracting new insights from your data. Within this paradigm you may run traditional analysis from either domain, extracting shortest paths and community groupings from the graph world, or detect trends and anomalies via time series.
However, where things get interesting is that these can be freely combined. This allows you to look for time-respecting paths through your data, model the evolution of communities over time or extract temporal patterns in the interactions between entities in the network. This can generate a deeper understanding within many different contexts, for instance: In social networks, we can better track the outreach of marketing campaigns or, contrarily, control the spread of fake news; In finance, it helps identify emerging markets and allows for a timely response to financial crime; In supply chains, it brings to light highly coupled dependencies and how service outages may impact delivery.
Within this talk we will introduce the Raphtory model for data ingestion and analysis. We shall peek under the hood to give an understanding of how the platform works. And finally, we will dive into recent use cases to give some inspiration as to where it may be applied.
Naomi Arnold is a postdoctoral research assistant at the Network Science Institute of Northeastern University in London. Her research focuses on making sense of large connected systems by representing them as temporal graphs and developing algorithms for their analysis. Most recently, this has been applied to using temporal motifs to understand cryptocurrency transaction networks and communication patterns in online social networks. She is also a developer for the FETA (Framework for Evolving Topology Analysis) and Raphtory open-source software for temporal network modelling and analysis, respectively. Previously, she completed a PhD and postdoctoral position in Computer Science at Queen Mary University of London.
Ben Steer is the CTO and Co-Founder at Pometry as well as an honorary researcher at Queen Mary University of London. His work in complex networks and distributed systems focuses on the development of Raphtory, an open-source tool for temporal graph analytics. His secondary focus as a board member of the Linked Data Benchmark Council is to define frameworks where different graph-based technologies can be fairly tested and compared and help researchers open new frontiers in high-performance graph data management.
Time: 15.00 – 16.00
Date: Tuesday 6 June
Location: I-X 5 Meeting Room, Level 5
Translation and Innovation Hub (I-HUB)
Imperial White City Campus
84 Wood Lane
London W12 7SL
If you would like to book on to this event please email Dr Hamed Haddadi, (firstname.lastname@example.org).