More and more Cultural Heritage (CH) data are available as Linked Open Data (LOD) in SPARQL endpoints. However, a recurring practical problem is how to make such data more accessible and easier to use for Digital Humanities (DH) research without requiring advanced SPARQL querying and programming skills. This talk presents and demonstrates how this challenge has been tackled in practice using the Sampo-UI framework [9, 12].
Sampo-UI is a JavaScript framework for creating web applications or “semantic portals” on top of SPARQL [4] endpoints. Sampo-UI is an open-source project and the code is available on GitHub¹. It has been used in more than 20 DH “Sampo portals” [6] in Finland with up to millions of users, covering various domains such as parliamentary data [8], nineteenth-century letters [7], and historical opera performances [1]. In recent years, it has also been used in multiple international projects in, for example, Norway, Belgium, the Netherlands, Latvia, and Switzerland [13, 3, 11, 10, 2]. Sampo-UI can be used to implement faceted search and browsing functions [5, 14], data analytic tools, and visualizations. While Sampo-UI has been used mainly in CH applications, it aims to be domain independent.
A new web application can be developed quickly by taking an existing Sampo-UI system as a starting point and 1) changing its SPARQL endpoint address, 2) adapting the SPARQL queries to the metadata model of the new application, and 3) configuring the available data analytic tools and visualizations. This requires only modest SPARQL and programming skills, while more experienced programmers can further extend the framework with novel components and functionalities. Based on the Sampo model [6], the framework separates the user interface from the underlying LOD service and uses only the SPARQL API to access the data. This means that Sampo-UI can be used to create applications for open SPARQL endpoints that are not controlled by the application developer [3].
A tool like Sampo-UI can help open CH data to the general public by lowering barriers to engaging with it. Faceted search and interactive visualizations can help all people interested in CH interact with the data without requiring deep technical expertise, such as writing your own SPARQL queries or creating visualizations with programming languages. DH researchers can also benefit from such applications as they facilitate iterative and exploratory approaches to analyzing data. Faceted search allows researchers to refine their searches step by step in an exploratory fashion, so that they do not necessarily need to know in advance what they want to find, but can instead discover interesting phenomena during the process itself. Being able to create visualizations quickly with only a few clicks also allows researchers to experiment with different ideas without having to commit a lot of effort.
¹Sampo-UI GitHub: https://github.com/SemanticComputing/sampo-ui