Knowledge Graph

I like being understood. Like most people, I prefer it when others get me. When I go back to Pietermaritzburg, am with my school and university friends, or when with family, it feels as if I am with people who get me. Know both me and my context.

I think that we make sense of the world by connecting data and arranging it into information entities. I feel knowledgeable when I can create a cohesive picture from the accessed data, information and entities. Jean (I think) does this in words. I usually represent this information entity visually with pictures, charts, a Johari Window, the dreaded pyramid or other familiar patterns. This can sometimes be a bit static and flat.

WORDLE

For people who do not do Wordle, this collection of multi coloured squares seems pretty meaningless. For those who understand the activity. They understand what this means.

⬜️🟨⬜️🟨🟨 🟩⬜️🟨⬜️🟨 🟩🟨⬜️🟨🟩 🟩🟩🟩🟩🟩

For those who do the Wordle regularly, the collection of daily guesses to arrive at the correct solution can be analysed. Grant Sanderson has done a lovely deep dive. With a subscription to NY Games, you too can buy data-driven insights to improve your Wordle strategy.

So familiarity with the game offers a context for understanding the system. The structure of the game enables programmers to create a clever machine that assists them. This ongoing collection of data and comparisons between all NY Times Wordle players allows for the formation of a knowledge graph. Which is explained far better than I can by Wey Gu

Most of the software systems I’ve used in education have focused on the storage and retrieval of data. The facilities for entering data are form fields or a grid. The result is database of rows and colomns. Data is extracted outof the rows and columns and represented in different ways to arrange it (table, bar chart, pyramid) so that it makes visual sense to others. But too often, we forget about the context around the data, or lack the insights about the people who are answering questions or collecting data.

A knowledge graph is another way of arranging data and information. Instead of holding data static in rows and columns, a knowledge graph organizes information in its natural form: a web of interconnected entities. The mapping of these entities — objects, events,
or concepts — and their relationships creates an interconnected schema. I’m still new at this and I don’t quite understand edges and nodes. That (hopefully will come).

A flexible schema behind a knowedge graph makes it possible to add new data and information and relationships between the data points can suddenly emerge. Patterns that might get lost easily in a traditional database become visible with a knowledge graph.

Magnus has shared an amazing lot of open source tools that make use of Wikidata. I think it’s his code and the thoughts about the pluriverse (thanks Michael) or transmodernism (thanks Nicholas) that have prompted this post.