Data dictionary
High-level description of the main fields used in SchoolSight's incident analysis schema. This is static content you can shape to match your technical documentation.
| Field | Type | Description |
|---|---|---|
id |
string | Unique identifier for the incident. |
risk |
string | Categorical risk band (e.g. Low, Medium, High). |
fusion_score |
float | Normalised 0–1 risk score combining media forensics, text analysis and contextual factors. |
deepfake.probability |
float | Model-estimated probability that the asset is synthetically generated or manipulated. |
text_analysis.toxicity_score |
float | Normalised toxicity score from the text pipeline (0–1). |
text_analysis.slang_terms |
array<string> | List of slang terms detected in the text, resolved against the SchoolSight lexicon. |
chain_of_custody.file_hash |
string | SHA-256 hash of the original file bytes used to verify integrity over time. |
chain_of_custody.first_seen_at |
datetime | Timestamp when the incident file was first ingested by SchoolSight. |
context.school_id |
string | Identifier for the school or campus associated with the incident. |
context.subject_key |
string | Pseudonymised key linking incidents to a subject timeline without storing names. |