Data Sources and Datasets

Collection of data from the ICPSR related to arts and culture. Includes information on arts participation, employment, demographics, and music.

The National Gallery of Art has started operating under an open data policy and provides access to files describing museum holdings, locations, and related organizations, people, and media outlets.

Collection of US federal data sources that have art-related variables and other notable datasets.

Offers both datasets and data "recipes" to create data from APIs, web scraping, or data mining techniques for an assortment of topics.

Links to some datasets about film and literature as well as politics, social media, and more.

List of unique datasets ranging from list of Dr. Who villains, Scottish witchcraft, and NYC popular dog names.

Database dedicated to documenting composers that are from historically excluded groups. There are databases for songs, choral , orchestra, and wind band works.

The New York Philharmonic Orchestra documenting performances from 1842 until present with information on pieces performed, conductor, players, soloists, date, time, and venue.

Offers some statistics and snapshots of data from the music industry. Unfortunately does not have a bulk download option.

Commonly used dataset containing audio features and metadata for over a million contemporary music tracks.

Link to an action recognition dataset from over 13,000 YouTube videos across 101 action categories.

From the popular app, a dataset of 50 million drawings across 345 categories has been compiled and made available for training neuro networks and AI.

Contains emotional reactions to art (mostly paintings) from WikiArt.org's collection. Contains information on how much participants liked the image, the painting's genre, and subject matter across several scenarios.

Files cataloging MoMA exhibits by artists and artworks.

Website listing access to art museum and collection datasets from several institutions such as the Tate Collection, MOMA, and the Metropolitan Museum of Art.

Open access to 3D point clouds and surfaces of historical sites around the world. Offers viewing on their website and public download.

Dataset containing metadata on over 200,000 English-language fiction volumes in HathiTrust.

.txtlab is a McGill University laboratory focused on cultural analytics. They provide access to several literature and text datasets used in their analyses including ones on contemporary novels, race and film, 20C poetry, subsets of the HathiTrust Library's holdings.