The Problem
When users approach large knowledge graphs, they encounter a fundamental tension between two needs: understanding what's possible within the graph's structure and finding the specific data they seek. Traditional interfaces force an uncomfortable choice. Either users must learn complex query languages like SPARQL or Cypher, memorizing schemas and syntax before they can ask even simple questions, or they're presented with node-link diagrams that collapse into illegible hairballs the moment the graph exceeds a few hundred nodes.
Both approaches fail at scale, but more importantly, both fail to recognize that query construction and schema learning are intertwined activities that should support each other.
The Pattern: Incremental Query Construction
Incremental query construction reimagines querying a graph via navigation. Rather than showcasing all potential data in a graph, Increment directs users through a guided exploration where each step forward reveals only the paths that actually exist in the data. Think of it like Just-in-Time scaffolding where the options at each query represent real structural possibilities within your knowledge graph.
Users build queries not by writing them wholesale but by making successive choices, with each selection naturally constraining what comes next. The interface becomes a map of possible queries where information scent guides users toward their goals without requiring them to understand the underlying graph structure beforehand.