Dataset v0.4.0 · schema 0.4 · 1895–2025

Data dictionary

Film Genealogy separates historical research records from visualization metadata. The schema makes the project’s claims inspectable without pretending that every edge has completed external scholarly review.

Cite this dataset

Kao, Ronnie. Film Genealogy: An Interactive Cinema Genealogy. Dataset v0.4.0, coverage 1895–2025, reviewed October 4, 2026.
https://tienhsiangkao.github.io/film-genealogy/

Last reviewed: 2026-10-04
Persistent identifier: none yet. The versioned web release is citable, but no DOI has been minted.

Downloads

JSON Nodes CSV Edges CSV Sources CSV

The JSON contains two explicit top-level layers: research and visualization.

Research layer

Node fields

id: Stable machine identifier within this dataset.
name: Public display label for the historical entity.
dateLabel: Human-facing historical date or range as displayed by the atlas.
startYear / endYear: Structured date anchors. End year remains null when the represented range is open or not yet normalized.
datePrecision: year, circa, range, decade_or_range, or open_ended.
dateBasis: Explains that the date represents emergence within this genealogy rather than birth date or a universal career start.
category: Editorial research lane: reality, narrative, formal, or production.
entityType: Structured ontology term such as person, movement, technology, production_system, practice, historical_current, or historical_concept.
role: Human-readable historical role.
idea: Compact statement of the node’s relevance to cinematic transmission.
description: Curatorial explanation of the node.
films: Representative works or transmission objects.

Research layer

Claim / edge fields

claimId: Stable identifier for an independently addressable historical claim.
a / b: Source and target node identifiers for the displayed directional relationship.
relationType: What historical relationship is being asserted: for example explicit influence, remake, movement practice, formal lineage, production model, or technological enablement.
evidenceType: Why the atlas is willing to draw the line: acknowledged, historical_reconstruction, or industrial_context.
displayCategory: Visual styling category retained for the map legend: acknowledged, lineage, or context.
claim: Curatorial proposition represented by the line.
sourceIds[]: Structured bibliography records supporting the claim.
sourceLocators[]: Claim-level page, passage, heading, timestamp, or an explicit locatorUnavailable reason. Null is preferred to invented precision.
assertedBy / assertedDate: Records who authored the curatorial assertion and when this schema version recorded it.
reviewStatus: Current editorial review state. v0.4 uses curatorial_unreviewed; it does not imply external scholarly review.
reviewedBy / reviewedDate: Reserved for real review events only; currently null.
counterSourceIds[]: Optional sources that complicate or contest the claim; currently empty unless evidence is actually identified.

v0.4 establishes claim-level provenance fields but does not fabricate page numbers, passage identifiers, reviewers, or counter-sources. Most current promoted edges explicitly state that a claim-level locator remains to be verified.

Research layer

Source fields

id: Stable source identifier used by edge claims.
sourceType: Structured class such as scholarly_monograph, primary_writing, interview, institutional_criticism, or institutional_reference.
authors[]: Named authors only when already verified in the dataset; otherwise empty rather than inferred.
title: Structured work/article title.
containerTitle / publisher: Publication or institution when known.
publicationYear / edition / volume / issue / pageRange: Bibliographic fields; null where not yet normalized.
doi / isbn: External publication identifiers when known; currently null unless verified.
url / archiveUrl / accessed: Web access metadata. Missing values are explicit rather than guessed.
label: Human-readable citation retained for interface display.

Research layer

Reading-route fields

id / label: Stable route identifier and public title.
thesis: The interpretive proposition of the route.
pathType: editorial_traversal: adjacency is pedagogical, not automatically causal.
steps[]: Ordered node stops. A transition either references a promoted transitionEdgeId or supplies a transitionNote explaining the editorial jump.

Visualization layer

Layout metadata

nodes[].x / y: Presentation coordinates. X is chronology-derived; Y is editorial and not a quantitative similarity measure.
dx / dy: Hand-authored collision/legibility offsets.
major: Visual prominence flag used by semantic zoom.
portrait: Wikimedia Commons filename used by the presentation layer; image rights remain source-specific.
chapter: Navigation-window assignment only; not a historical claim that one movement replaces another.
currents: Interpretive visual ranges with start/end years, vertical placement, height and tone.
chapters / lanes: Navigation and editorial spatial grammar used by the atlas interface.

Selection protocol

A node is included when it represents a historically significant transmission that can be supported by at least one high-quality source.
A node should materially connect parts of the atlas or clarify one of the project’s research questions.
Inclusion does not signify rank, canon status, or overall historical importance.
A promoted edge requires direct acknowledgement, defensible historical/formal transmission, or a material production/technology relationship. Similarity alone is insufficient.

Research backlog

African cinemas · queued
A coherent branch with multiple internal relationships and claim-level sources; not a token attachment to an existing European line.
Latin American cinemas · partial
Expand beyond the current Third Cinema abstraction with researched national/transnational branches and specific claims.
Arab cinemas · queued
A scoped corpus, explicit historiographic framing and claim-level sources.
Indigenous cinemas · queued
Community-aware and scholarly framing established before graph inclusion.
Animation · queued
Adapt the taxonomy to studios, techniques, labor systems, schools and technological lineages rather than auteur tokens.
Postwar experimental / avant-garde cinema · queued
A branch with its own institutions, practices and transmissions.
Documentary lineages · partial
Expand beyond broad nonfiction currents with specific documentary practices and source-backed transmissions.
Queer film histories · queued
Multiple researched lineages rather than representative token inclusion.
Below-the-line labor · queued
Schema and evidence support editors, cinematographers, sound, performers, production design and producers as historical agents.

Rights by layer

Original dataset / editorial metadata

The author has not yet designated a formal reuse license for the original Film Genealogy dataset/editorial metadata.

Website / source code

Website/source-code reuse is governed separately from the research dataset.

Portraits

Portrait images remain governed by the license and attribution terms on their Wikimedia Commons source pages.

Cited sources

Books, articles, interviews and web sources are cited as evidence and are not relicensed by this dataset.

A formal open-data license has deliberately not been chosen automatically. That is an authorial/legal reuse decision, not a field that should be filled merely to make the schema look complete.