Data Science & Analytics

CineJaal

Interactive cultural archive for exploring the interconnected universe of Hindi cinema across films, artists, studios, music, and franchises.

Next.jsTypeScriptGraph VisualizationPythonWikidata

Overview

CineJaal is a browser-based Hindi cinema archive connecting films, actors, directors, music directors, singers, production houses, and franchises. Its progressive force-directed canvas stays synchronized with a 1970–2026 timeline, supports shortest-path exploration, and pairs sourced relationship data with an original deterministic poster system.

Highlights

  • Archives 500 real Hindi-language films released between 1970 and 2026.
  • Connects 2,096 deduplicated entities through 4,953 sourced relationships.
  • Combines a progressive force-directed canvas with a synchronized historical timeline.
  • Supports shortest-path exploration between people, films, studios, and franchises.
  • Runs from bundled JSON at runtime, avoiding upstream availability and latency dependencies.

Archive & Graph Experience

The archive presents Hindi cinema as a connected network rather than a catalog of isolated titles. Visitors can move through a force-directed canvas while the synchronized timeline grounds each discovery in its release era.

  • Covers films, performers, directors, music directors, singers, production houses, and source-backed franchises.
  • Progressively reveals the graph so a large network stays navigable in the browser.
  • Shortest-path exploration explains how any two entities are connected.

Data Pipeline & Provenance

  • Build scripts can reuse cached sources, reproduce the committed snapshot offline, or refresh upstream inputs.
  • The deployed application reads bundled JSON and makes no upstream data calls at runtime.
  • Wikidata supplies CC0 structured data while IMDb fields follow its non-commercial dataset terms.
  • Validation scripts check the generated archive before it is shipped.

Original Visual System

  • Deterministic CSS geometry creates a consistent poster identity from factual movie metadata.
  • No theatrical posters, film stills, celebrity likenesses, or third-party images are bundled.
  • The visual language keeps the archive expressive while respecting source and licensing boundaries.