Storytelling with Media Suite Data Stories
By Roeland Ordelman, Netherlands Institute for Sound and Vision
Media Suite – Data Stories serves as an online platform for storytelling based on data. It presents narratives interspersed with visualisations, gleaned from interesting facts and figures that are extracted from large archives of AV media. Such narratives can be part of an academic endeavor, an investigative journalism report, the outcome of a datathon, or assembled by archivists inspired by events in the news. The topic can be anything from a story looking back over the 15 year run of a highly successful TV programme, to a semantic analysis of the ambiguity of information about the coronavirus as reported in the media. The Media Suite offers the core functionalities needed for this. Tools for distant and close reading, various advanced options for data analytics, and a workspace for conveniently saving work in projects.
The Media Suite is a virtual research environment. Data stories are for a general audience, but the Media Suite is designed to be used by scholars and investigative journalists who are interested in a data science approach, using AV collections and related multimedia data such as TV program guides, film schedule information, viewer ratings, scripts, posters and photos. The data come from Dutch institutes such as the Netherlands Institute for Sound and Vision (NISV), the Eye Film Institute, and the Royal Library. The Media Suite is part of what we sometimes playfully refer to as a storytelling superhub. A node in the nascent Research & Heritage data infrastructure in the Netherlands, initiated by the CLARIAH project for Digital Humanities and Social Sciences and supported by the Dutch Digital Heritage Network. NISV is responsible for the development and long term maintenance of the Media Suite, in close collaboration with partners within the infrastructure.
We are experimenting with new ways to open up, explore and exploit the multifaceted richness of large multimedia archives, focussing on disclosure, access, analysis, interfacing, visualisation, storytelling and community involvement.
The Media Suite incorporates federated authentication services. Some of the data is open access and can be used by anyone. But to enable access to copyright-protected or privacy-sensitive data for academic research or journalism, we can tailor access to groups (academic researchers, collection owners), and individuals (journalists, researcher without a university affiliation, citizen scientists). Access rights are controlled by the collection owners themselves via collective agreements and by issuing individual accounts.
The Media Suite uses artificial intelligence for large scale automatic metadata extraction such as speech recognition, computer vision and natural language processing. This boosts its capabilities for producing insights for data stories. For brave users with a do-it-yourself attitude, the Media Suite facilitates programming your own data insight discovery via APIs and a programming environment (Jupyter Notebooks). This leads to new investigative possibilities and represents the middle ground between being a data novice and a fully fledged programmer.
The Media Suite is an innovation lab. We are experimenting with new ways to open up, explore and exploit the multifaceted richness of large multimedia archives, focussing on disclosure, access, analysis, interfacing, visualisation, storytelling and community involvement. We are in a constant process of reflection and agile co-development with our users, striving towards a better user experience. For this, we are delighted to be able to collaborate with our partners in academia and higher education in a variety of research and innovation projects funded by national and European funding schemes.