What is the OldVis project?
The OldVis project is a digital collection dedicated to historical visualizations. The digital collection is built by assembling information related to old visualizations, which are scattered across existing digital libraries such as the David Rumsey Map Collection and the Internet Archive. As of August 17, 2023, it has indexed over 13,000 historical visualizations published before 1950.
Mission
Historical visualizations have played crucial roles in science and communication. For example, John Snow's Cholera Map (1854) laid the foundation for epidemiology. These visualizations represent important milestones in the history of data visualization.
While historical visualizations are valuable records that offer insights into how data was communicated and interpreted across different eras, they have remained scattered across various digital libraries, making systematic study difficult. The OldVis project aims to address this gap in digital humanities and visualization research: the lack of a large-scale, dedicated collection of historical visualizations.
By creating a comprehensive digital collection, the OldVis project aims to:
- Gather scattered historical visualizations into a unified, searchable collection.
- Make historical visualization materials easy to search, browse, and reuse.
- Enable systematic analysis of visualization techniques, design patterns, and their evolution over time.
Components
The OldVis project consists of several components:
| Repository | Description |
|---|---|
| gallery | An online gallery of old visualizations (details). |
| oldvis_dataset | A Python package to download the processed catalog (details). |
| dataset | Data curation scripts in Python and the built dataset (details). |
| image-taxonomy | A taxonomy of old visualizations (details). |
| libquery | A Python package to query digital libraries (details). |
| libquery_extensions | A Python package that extends libquery (details). |
| libprocess | A Python package to process query results from libquery (details). |
| libprocess_extensions | A Python package to process query results from libquery_extensions (details). |
| image classification labeler | An interactive data labeling system for image classification (details). |
| image segmentation labeler | An interactive data labeling system for image segmentation (details). |
| image taxonomy labeler | An interactive data labeling system for image taxonomy labeling (details). |
| docs | This documentation website. |
This website maps the collection and the tools. Each repository's README and docs go into more detail.
Stack
The stack has four layers, from data collection up to the gallery:
- Data collection: Fetch metadata and images with libquery and libquery_extensions, then process them with libprocess and libprocess_extensions.
- Data labeling: Label images with the image classification labeler, image segmentation labeler, and image taxonomy labeler.
- Dataset: Curation scripts and the built catalog live in dataset. oldvis_dataset downloads that catalog.
- Gallery: The gallery is the public browser for metadata and images.
Copyright
Notice
We do not hold the copyright to the images of old visualizations indexed by the OldVis project. We have stored the rights information obtained from the data sources when available. When the rights information requires attribution (e.g., when the image is licensed under CC BY-NC-SA 3.0), you should attribute the original data source instead of us. If you want to redistribute the images, you should check whether they are in the public domain according to the copyright lengths and regulations in their country.