OldVisOnline: Curating a Dataset of Historical Visualizations

Area of Science:

• Digital Humanities
• Information Visualization
• Historical Research

Background:

• Digitization efforts have made historical visualizations accessible online.
• Existing historical visualizations are fragmented across various digital libraries, hindering research.
• There is a lack of a dedicated, large-scale digital collection for historical visualizations.

Purpose of the Study:

• To curate the first large-scale dataset of historical visualizations.
• To develop a system for accessing and interacting with this dataset.
• To explore new research avenues in visualization and history.

Main Methods:

• Curated a dataset of 13,000 historical visualization images from seven digital libraries.
• Developed a workflow for scraping and processing heterogeneous metadata.
• Implemented a semi-automatic labeling approach to identify visualizations.
• Built the OldVisOnline system for browsing and labeling.

Main Results:

• Established a comprehensive dataset of historical visualizations.
• Created a functional system (OldVisOnline) for dataset access and interaction.
• Demonstrated the feasibility of large-scale historical visualization data curation.

Conclusions:

• The curated dataset and OldVisOnline system provide a valuable resource for visualization and history research.
• The dataset enables new research opportunities, including textual criticism of historical visualizations.
• Recommendations are provided for future dataset enhancement and expansion.