Tutorials showing how to use OpenSanctions data in use cases.
1 Like
Knowledge Graph Conference 2025 will be held on May 5-9 in NYC, at Cornell Tech on the beautiful grounds of Roosevelt Island.
Prashanth Rao and I will lead a tutorial workshop on Monday, May 5, 11:00-12:30 US Eastern
“Creating High-Quality Knowledge Graphs From Structured and Unstructured Data”
We’ll work through examples using slices of OpenSanctions and Open Ownership datasets:
- Run Senzing entity resolution to identify entities and relations among them.
- Build a graph in KùzuDB using Polars.
- Show interactive visualizations in Jupyter notebooks with yWorks, and with the Kùzu Explorer UI.
There’s a helpful combination of open source tooling in the second and third steps: KùzuDB embedded graph data (can run on your smartphone!), Polars for scalable dataframes, yWorks visualizations within Jupyter notebooks, etc.
The materials are available on GitHub (to be shared at the workshop), which you can run on a laptop. This workshop will be held in-person at Cornell Tech, and also available for remote attendees.
PS: message me if you want a discount code for registration
2 Likes
Hope the workshop goes really well next week, @paco. Always appreciate the great work you are doing to show off use cases for Open Ownership data.
I had a good call with Prashanth and the Kuzu team a couple of months ago so will be keen to check out the materials once they are released on Github.
2 Likes
Thank you Stephen!
Here’s our repo for the tutorial at KGC next week:
Notably, here’s an integration of KùzuDB + yWorks + Jupyter showing an open source stack for graph analytics and interactive visualization, with each component running as in-process libraries. In other words, these follow more of a DuckDB pattern and can be launched simply as Python packages, not as platforms which require ops overhead, stateful deployments, etc. KùzuDB provides both graph and vector data, which is not memory-bound.
Any feedback is much appreciated!
2 Likes