Datasetpublic

wonwooyoon/metatake-filming-locations

Metatake Film Filming-Locations Dataset A structured, geocoded map of where films were shot and set — 17,341 locations across 1,917 films and 130 countries, each with coordinates. A layer that mostly does not exist as open data elsewhere: it distinguishes the filmed location from the narrative setting, and records each place's role in the film. Curated and published by Metatake — an independent film-criticism platform. See the full per-film pages, e.g. Mulholland Drive.… See the full description on the dataset page: https://huggingface.co/datasets/wonwooyoon/metatake-filming-locations.

sourceHugging Facecc-by-4.0updated 2mo agoView on Hugging Face

0likes14downloads

Dataset Card

Metatake Film Filming-Locations Dataset

A structured, geocoded map of where films were shot and set — 17,341 locations across 1,917 films and 130 countries, each with coordinates. A layer that mostly does not exist as open data elsewhere: it distinguishes the filmed location from the narrative setting, and records each place's role in the film.

Curated and published by [Metatake](https://metatake.net) — an independent film-criticism platform. See the full per-film pages, e.g. Mulholland Drive.

Files

File	Rows	Notes
`metatake-filming-locations.jsonl`	17,341	one JSON object per location (recommended)
`metatake-filming-locations.csv`	17,341	same data, CSV
`STATS.json`	—	summary counts

Fields

field	description
`film_slug`, `film_title`, `film_year`	the film (slug is its metatake.net id)
`imdb_id`, `tmdb_id`	cross-ids for joining to other datasets
`name`	the place (e.g. "Palace Theatre, 630 South Broadway, Los Angeles, USA")
`role`	its role in the film (scene / production role)
`narrative_setting`	what it stands in for in the story (may differ from where it was filmed)
`layer`	`filmed` (a real shooting location) or `setting` (the fictional place depicted)
`kind`	granularity: `city`, `region`, `landmark`, …
`country`	country of the location
`lat`, `lng`	WGS-84 coordinates (100% of rows)
`precision`	geocoding precision, where recorded
`confidence`	Metatake's confidence in the mapping, where recorded

How it was built

Locations are extracted per film from public reference sources and Metatake's own close-reading pass, then geocoded and reconciled. Two passes separate filmed from setting. Method notes: https://metatake.net/methodology#locations . This is curated interpretive data, not a comprehensive shooting-permit registry — coverage skews to films Metatake has read closely.

Usage

python

from datasets import load_dataset
ds = load_dataset("<owner>/metatake-filming-locations")  # after upload
# or read the file directly:
import pandas as pd
df = pd.read_json("metatake-filming-locations.jsonl", lines=True)

License & citation

Licensed CC BY-NC 4.0 — free to use, share, and adapt with attribution to Metatake, non-commercial. Attribution is the whole deal (the legal form of the "cite us" line every Metatake surface carries).

bibtex

@misc{metatake_filming_locations,
  title  = {Metatake Film Filming-Locations Dataset},
  author = {Metatake},
  year   = {2026},
  url    = {https://metatake.net},
  note   = {CC BY-NC 4.0}
}

Live API for the same data (single film / by country): https://metatake.net/api