Movies, TV series, directors, actors, writers and genres from the IMDB non-commercial dataset — modelled as a graph. Questions that feel natural to a film buff ("who collaborates with whom?", "which director keeps picking the same actors?", "how did ratings shift decade to decade?") become single Cypher traversals instead of multi-table SQL joins.
cd case_studies/imdb-movies && ./run.sh # validate every query against the snapshot
RECORD=1 ./run.sh # also regenerate demo.gifScale: 1,940,360 nodes · 2,634,125 edges (imported from a 49.8 MB snapshot in seconds)
| Node label | Count | Key properties |
|---|---|---|
| Movie | 49,632 | tconst, title, year, runtime_minutes, title_type |
| Person | 293,550 | nconst, name, birth_year, death_year |
| AlternateTitle | 1,517,802 | title, region, language |
| Series | 14,858 | tconst, title, year, end_year |
| Rating | 64,490 | average_rating, num_votes |
| Genre | 28 | name |
Relationships (7):
| Relationship | Pattern | Count |
|---|---|---|
HAS_ALTERNATE_TITLE |
Movie/Series → AlternateTitle | ~1.5M |
ACTED_IN |
Person → Movie/Series | ~530K |
DIRECTED |
Person → Movie/Series | ~80K |
WROTE |
Person → Movie/Series | ~90K |
PRODUCED |
Person → Movie/Series | ~270K |
HAS_RATING |
Movie/Series → Rating | ~64K |
HAS_GENRE |
Movie → Genre | ~120K |
The director–actor collaboration network — (:Person)-[:DIRECTED]->(:Movie)<-[:ACTED_IN]-(:Person) — is the star: a bipartite graph that turns "power pairs" into a first-class queryable structure.
See queries.cypher. The narrative: crowd-favourite films →
most prolific directors → genre dominance by audience size → director–actor power
pairs (the multi-hop graph-native question) → actors with widest director reach →
cinema by decade → top TV series → writers per director.
Source: IMDB Non-Commercial Datasets
— title.basics.tsv.gz, title.ratings.tsv.gz, name.basics.tsv.gz, title.principals.tsv.gz, title.akas.tsv.gz.
License: IMDB Non-Commercial Use Only. This data is provided for personal, non-commercial use only. See IMDB terms for full conditions.
Snapshot built by examples/imdb_loader.rs
with --min-votes 1000 --min-votes-series 500 --min-year 1950 --akas title.akas.tsv.gz and published as
imdb.sgsnap on release
kg-snapshots-v8
(sha256 pinned in case.env).