ARTigo: Social Image Tagging (Aggregated Data)

ARTigo (https://www.artigo.org/) is a Citizen Science project that has been jointly developed at the Institute for Art History and the Institute for Informatics at Ludwig Maximilian University of Munich since 2010. It enables participants to engage in the tagging of artworks, thus fostering knowledge accumulation and democratizing access to a traditionally elitist field. ARTigo is built as an interactive web application that offers Games With a Purpose: in them, players are presented with an image – and then challenged to communicate with one another using visual or textual annotations within a given time. Through this playful approach, the project aims to inspire greater appreciation for art and draw new audiences to museums and archives. It streamlines the discoverability of art-historical images, while promoting inclusivity, effective communication, and collaborative research practices. The project’s data are freely available to the wider research community for novel scientific investigations.

File structure

The dataset is provided in a .jsonl file format, with each line representing a single image and its associated metadata. The images themselves are provided separately in a .zip file.

1. data.jsonl: Each line in the .jsonl file represents a single image and its associated metadata, and has the following key-value pairs:

   • id: a unique identifier for the image;
   • hash_id: a unique identifier for the image based on its content (e.g., image hash);
   • titles: a list of titles associated with the image, with each title having the following key-value pairs:
     • id: a unique identifier for the title;
     • name: the name of the title;
   • creators: a list of creators associated with the image, with each creator having the following key-value pairs:
     • id: a unique identifier for the creator;
     • name: the name of the creator;
   • location: the location associated with the image;
   • institution: the institution that holds the image;
   • source: information about the source of the image, with the following key-value pairs:
     • id: a unique identifier for the source;
     • name: the name of the source;
     • url: the URL of the source;
   • tags: a list of tags associated with the image, with each tag having the following key-value pairs:
     • id: a unique identifier for the tag;
     • name: the name of the tag;
     • language: the language of the tag (if available);
     • count: the number of times the tag has been applied to the image;

   • path: the path to the image file.

2. media.zip: The images themselves are stored in a .zip file. Each image is stored in a folder named after the first two characters of its hash_id. Within this folder, there is a sub-folder named after the next two characters of the hash_id. The image file itself is stored within that sub-folder and is named with the complete hash_id and .jpg file extension. The folder structure within the .zip file thus is as follows:

   root
|
├── f4
|   └── 22
|       └── f42236be6580338e9b98b8e00c0f4e49.jpg
├── 4c
|   └── d3
|       └── 4cd3f476b14abfcb2a91e6c8f2d356f6.jpg
└── ...

Terms of use

The data are provided “as is,” without any warranties of any kind. They are provided under the Creative Commons Attribution-ShareAlike 4.0 International license, and are updated monthly, so users can be confident they are accessing the most up-to-date information.