Sparker
A SEN observation tracker that uses Neo4j to surface correlations between student behaviours, built for special educational needs facilitators.
About
An observation tracker for special educational needs facilitators, built graph-native on Neo4j. The data model treats each observation as a node linked to a student, a field definition and a value, so finding correlations becomes a natural Cypher traversal rather than a complex SQL join. A flexible user-defined field system (text, number, boolean, tags, datetime) adapts to any school's vocabulary, and auto-detected co-occurrence correlations surface in per-student insight reports and a global feed ranked by strength. The SvelteKit code keeps a clean server/load separation with all Neo4j access isolated behind typed loaders. This prototype leaves temporal patterns, trigger analysis, multi-user auth, and richer visualisation on the roadmap.
Highlights
- Graph-native data model: (Student)<-[:ABOUT]-(Observation)-[:HAS_FIELD {value}]->(FieldDefinition). Correlation discovery as a Cypher traversal.
- Flexible user-defined field system: text, number, boolean, tags, datetime. Adapts to any school's observation vocabulary.
- Auto-detected co-occurrence correlations surfaced in per-student insight reports and a global insight feed ranked by correlation strength.
- Clean SvelteKit server/load separation: all Neo4j access isolated in src/lib/server/ behind typed +page.server.ts loaders.
- Neo4j query patterns kept in src/lib/server/ with typed loaders; the graph traversal for co-occurrence is a two-hop match across shared observation fields.
By the numbers
- Commits
- 11
- Lines added
- +789
- Lines removed
- −8
- Source files
- 1,732