Summary

This page provides key metrics about our knowledge graph (KG), including its size, density, and connectivity patterns, with a focus on how nodes from our disease and drug lists are connected within the graph.

Graph size

9.34M
nodes

68.46M
edges

Graph density

7.3
edges per node on average

4.9
edges per node when excluding the top 1,000 most connected nodes

Upstream data sources

Epistemic Robustness

Epistemic Score

The Epistemic Score summarizes provenance quality across the graph by averaging values assigned to each edge's knowledge level and agent type.

A more positive average reflects stronger overall provenance, while a more negative average indicates weaker or more speculative knowledge across the graph.

0.29
Epistemic Score

68.46M
edges used in calculation

46.38M
edges with missing provenance

“Missing provenance” includes edges where both Knowledge Level and Agent Type are "Not Provided" or not present.

Knowledge Level

Indicates how strong or certain a statement is—ranging from direct assertions and logical entailments to predictions and statistical associations.

A more positive average reflects greater epistemic confidence, while a more negative average indicates weaker or more speculative knowledge.

0.30
Average Knowledge Level

68.46M
edges used in calculation

Agent Type

Agent Type describes the origin of an edge in terms of how the knowledge was generated—ranging from direct human assertions to automated text mining.

A more positive average reflects greater human involvement, while a more negative average indicates greater reliance on automated or speculative sources.

0.29
Average Agent Type

68.46M
edges used in calculation

Disease list nodes connections

308.2
edges per disease node on average

Drug list nodes connections

1,704.5
edges per drug node on average