View this email in your browser
Topological Data Analysis and Knowledge Graphs
In mathematics, topology identifies the basic properties of a geometric object that remain constant even if external features change (e.g., the object is bent or twisted into a different shape). Homology identifies relationships between objects. Topological data analysis (TDA) is a way of digesting data into basic components and relating data sets, both of which are essential to creating knowledge graphs.

Christopher Severs, who helped launch Knowledge Graph work at eBay, was the first person who pointed me toward TDA. Now there's a small but growing constellation of advanced mathematics and artificial intelligence use cases in homology. In other words, can you apply relatively advanced linear algebra and topology to "scale" a graph from "fine-grain detail" to "big picture" while preserving its structure?

If so, then some big data use cases can leverage graph techniques. For example, DataRefiner is a spin-off from Google's internal tools for customer segmentation. Edward Kibardin, CEO and founder of DataRefiner, explains that the product can combine data from diverse sources (user activity, sensors, text) to create a data map. That's an interesting use of topological data analysis (TDA) to determine graph structure within data. For instance, data from user interactions with apps, texts from user comments, and survey responses could be aggregated and used for marketing campaigns. Other possibilities include using data from financial transactions for anti-money laundering, or data collected from hardware to improve performance.

Symphony AyasdiAI (funded by friends at DCVC) is another product in this sphere, with applications in financial services, pharmaceutical research and government. 

One of the general ideas is "can you separate the foreground (structure) from the background (noise)" in a mountain of data, and begin to surface relations (graph structures). Admittedly, Ayasdi and others seem to have pivoted toward specific customer verticals. Even so, it's a general category of work where graph visualization is the thin edge of the wedge for relatively advanced work that heads toward constructing graphs from data.

-Paco Nathan
Graph Visualizations
Election 2020 Interference
This screenshot is from an interactive graph prepared by Atlantic Council.
It shows the number of Facebook and Twitter postings with misinformation intended to influence the election and countries from which the postings originated.
Security Incident and Event Management
Above is an example of Leo Meyerovich's Graphistry being used for Security Incident and Event Management (SIEM). The graph is a playbook showing event progression, linkages, scope, root cause and more.
Product Orders and Related Elements
This graph was shared by Juan Sequeda of Grafo, a product that simplifies creation of knowledge graphs. 
Twitter Data
Over at NodeXL, Marc Smith has a number of interesting projects underway. The graph represents a network of 42 Twitter users whose tweets over a two-week period contained "lrainie", or who were replied to or mentioned in those tweets.
Financial Transaction Data
This graph was created with Graphistry (Thanks to CEO Leo Meyerovich for the share on LinkedIn.) It tracks Suspicious Activity Reports (SARs) of transactions between banks for purposes of anti-money laundering (AML).
Educational Institutions
Mathieu Bastian's Gephi is an open-source application that creates graphs for use as a complement to traditional statistical analysis. The graph above shows connections between various European universities. Gephi users can customize colors, size or labels to more clearly illustrate relationships and patterns.
Expanded Star Wars Universe of Characters
Kirell Benzi is a data scientist and artist who will be a featured speaker at KGC’s Knowledge Connexions 2020 event. He and a colleague at École Polytechnique Fédérale de Lausanne (EPFL) created this graph showing the connections between the 7,563 main characters in the Expanded Star Wars Universe based on data compiled from web pages.
Human Resource Management
Dataveyes CEO Caroline Goulard shared this image from a project that aims to foster better HR strategies by visualizing the organization, its skill sets and career pathways as big graphs. This image shows career options at a company and how it is possible to progress from one job into other roles.
The SWD Community is a free online destination for honing data visualization and storytelling skills. The group has members all around the world, who can access practice exercises and participate in a monthly challenge, as well as discuss issues on a blog and give each other feedback on their efforts. 


Let us know what you're working on in Slack.
Copyright © 2020 Knowledge Graphs Conference, All rights reserved.

Want to change how you receive these emails?
You can update your preferences or unsubscribe from this list.

Email Marketing Powered by Mailchimp