DropTheHQ commited on
Commit
2de9910
·
verified ·
1 Parent(s): e5a489f

Add dataset card with full description

Browse files
Files changed (1) hide show
  1. README.md +92 -0
README.md ADDED
@@ -0,0 +1,92 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: mit
3
+ task_categories:
4
+ - text-classification
5
+ - feature-extraction
6
+ - zero-shot-classification
7
+ language:
8
+ - en
9
+ size_categories:
10
+ - 1M<n<10M
11
+ tags:
12
+ - knowledge-graph
13
+ - entertainment
14
+ - movies
15
+ - series
16
+ - people
17
+ - companies
18
+ - cryptocurrencies
19
+ pretty_name: DropThe Entertainment Knowledge Graph
20
+ ---
21
+
22
+ # DropThe Entertainment Knowledge Graph
23
+
24
+ A large-scale, open knowledge graph covering **1.8 million entities** and **2.9 million connections** across entertainment, media, finance, and technology. Built and maintained by [DropThe](https://dropthe.org), a data utility media network that organizes the world's entertainment and financial data into structured, queryable formats.
25
+
26
+ ## Dataset Description
27
+
28
+ This dataset represents a snapshot of the DropThe entity database, one of the most comprehensive open entertainment knowledge graphs available. Each entity is enriched with structured metadata, cross-references, and relationship links that connect people to their work, companies to their products, and media to its creators.
29
+
30
+ ### Entity Types
31
+
32
+ | Type | Count | Description |
33
+ |------|-------|-------------|
34
+ | Movies | 52,000+ | Feature films with cast, crew, ratings, streaming availability |
35
+ | Series | 14,000+ | TV shows and streaming originals with episode-level data |
36
+ | People | 1,700,000+ | Actors, directors, producers, executives, public figures |
37
+ | Companies | 8,000+ | Studios, distributors, production companies, tech firms |
38
+ | Cryptocurrencies | 12,000+ | Tokens and coins with market data and project metadata |
39
+
40
+ ### Data Fields
41
+
42
+ Each entity record includes:
43
+
44
+ - **Core identifiers**: Unique ID, canonical name, URL slug, entity type
45
+ - **Structured metadata**: Birth/death dates, nationalities, genres, ratings (stored as JSONB)
46
+ - **Popularity scoring**: Composite score derived from search volume, social mentions, and editorial signals
47
+ - **Tier classification**: S/A/B/C/D ranking based on cultural significance and data completeness
48
+ - **Temporal markers**: Creation date, last enrichment date, data freshness indicators
49
+
50
+ ### Relationship Graph
51
+
52
+ The 2.9 million connections in `entity_links` encode typed, directional relationships:
53
+
54
+ - `acted_in` / `directed` / `produced` -- People to Movies/Series
55
+ - `works_at` / `founded` -- People to Companies
56
+ - `distributed_by` -- Movies to Companies
57
+ - `related_to` -- Thematic and contextual associations
58
+ - `traded_on` -- Cryptocurrencies to Exchanges
59
+
60
+ ## Intended Use
61
+
62
+ This dataset is designed for researchers and developers working on:
63
+
64
+ - **Recommendation systems** -- Use entity relationships to build collaborative and content-based filters
65
+ - **Knowledge graph completion** -- Predict missing links between entities using graph neural networks
66
+ - **Named entity recognition** -- Train NER models on the 80,000+ aliases table
67
+ - **Entertainment analytics** -- Study trends in movie production, streaming availability, and talent networks
68
+
69
+ ## Methodology
70
+
71
+ The data is collected through a multi-stage pipeline documented on the [DropThe methodology page](https://dropthe.org/good/methodology/). Sources include public APIs (TMDB, CoinGecko, Wikidata), structured web extraction, and editorial enrichment. Every entity passes through validation checks for duplicate detection, data type conformity, and referential integrity before entering the production graph.
72
+
73
+ The full knowledge graph powers the entity pages on [DropThe](https://dropthe.org), where users can explore interconnected data across verticals including movies, series, crypto, companies, and people.
74
+
75
+ ## Sample
76
+
77
+ The included `sample_entities.csv` contains 20 representative entities across all five types, with core fields and a subset of metadata. The full dataset is available via the DropThe API and [data portal](https://dropthe.org/data/).
78
+
79
+ ## License
80
+
81
+ MIT License. Attribution appreciated but not required.
82
+
83
+ ## Citation
84
+
85
+ ```
86
+ @dataset{dropthe_entertainment_2026,
87
+ title={DropThe Entertainment Knowledge Graph},
88
+ author={DropThe},
89
+ year={2026},
90
+ url={https://dropthe.org}
91
+ }
92
+ ```