Embedded Research Projects 2012

For the period of 2012-2013 the UvA, VU and KNAW funded projects for embedded researchers in the area of digital humanities. The funding for the projects was limited to 30K euro and involved private partners or other 3rd parties.

UvA projects

  • Evaluating the Humanities [proposal and presentation] Across a number of research subjects in the arts and humanities, the book or monograph is the sine qua non of scholarly output.  Often it is a strong requirement for academic promotion and tenure.  When writing a book, the humanities scholar is likely to give serious consideration to the university or academic press with whom (s)he would like to publish.  One of the deciding factors will be the ‘prestige’ of the press.   In the end, what is important to the humanities scholar “is that the resultant publication from a prestigious university press is read, reviewed, and cited; it is purchased by libraries; it is taken seriously” (Gump, 2006, p. 76). Publisher ‘quality’ has been assessed by survey studies and it has also been related to library catalog holdings.  The concept of ‘prestige’ is typically associated with the selection of a university press for a new book.  Often it is also linked to the age of the press and its original mission, including the rank of its housing institution.  University press prestige is a well-discussed topic; however research has yet to be done to monitor or indicate this concept using bibliometric methods.  Due to the importance that humanities scholars place on choosing a press, one would expect to see the majority of citations to books published by those considered most prestigious.  Where books appear in the cited reference lists of journal articles, our main focus is to investigate whether or not certain university or academic presses are cited more often than others. One drawback pertaining to a study of cited books relates to the structure of current citation indices.  Both Thomson Reuter’s Arts and Humanities Journal Citation Index and Elsevier’s SciVerse Scopus include books in tagged reference lists of articles; however, each book lacks a distinct identification code within the index itself.  Here, we are working with a unit of information known as a reference ‘string’; hence, for millions of reference strings that can be identified as books (i.e., across the subjects of history and literature), we will develop a unique dataset that links cited books to publisher data found in worldwide academic library catalogs.
    Alesia Zuccala (principle researcher, UvA), Raf Guns (research assistant, UvA-University of Antwerp), Henk Moed and Gali Halevi (Elsevier Scopus, private partner), Roberto Cornaccia (Spinque, private partner).
  • Music Similarity [proposal, poster presentation, sheets presentation and publication] When is a piece of music similar to another piece of music? This is the case when two musical pieces belong to the same genre (eg Jazz or Rock), for example, or if the melodies are very similar. Can we measure this? In other words, can we have a computer compare those pieces and give a measure of equality as a result? This question is central to this project. Several properties of a piece of music are associated with the notion of musical similarity, for example, tempo, rhythm, meter, instrumentation and pitch contour. These properties, describing a part of the total similarity, are of importance for automatically finding a certain piece of music in a database. Using the division of musical similarity into parts, we can examine to what extent these sub-similarities contribute to categorization of style or genre. This brings us closer to answering the scientific question: What makes a piece of music similar to another piece of music? The more practical goal of this project is to build an application that finds, based on a particular piece of music, other music has a similar tempo / rhythm / timbre / ..? The resulting music-recommendation app will be used for many purposes such as: jogging app, DJ applications, music discovery app, and so on. One of the key features of this app will be that the user has control over its search criteria. He can search for music with similar tempo and rhythm, or music with similar instrumentation or search other preferred combinations. So the user can navigate along different axes of similarity, depending on the similarity which is desired at that moment. The research focus is therefore on interpreting and combining low-level features by an app-user to malleable meaningful “meta-features’.
    Aline Honingh (principle researcher, UvA), Bruno Rocha (research assistant), Victor Bergen Henegouwen (Elephantcandy, private partner), and Niels Bogaards (Elephantcandy, private partner).
  • Emotions in film [proposal in Dutch, presentation and draft follow-up research in Dutch] How can we detect emotions in film? In this project, a database (which includes 1000 movies) of emotion-tagged film shots is being created, based on earlier research by Professor Ed Tan (Professor of Communication Sciences) and Prof. Patricia Pisters (Professor of Film Studies).
    Ed Tan (Professor of Communication Sciences) and Patricia Pisters (Professor of Film Studies), VicarVision (private partner) and Hanneke Koster and Thuong Minh Bui (research assistants).
  • ExPoSe  [proposal , presentation and publication]
    Parliamentary proceedings and other transcripts of meetings are a common document genre characterized by a complex narrative structure. The essence is not only what is said, but also by who and to whom, and why. Standard search tools are tailored to the topical relevance of the content, focusing exclusively on the “what” is said. Modern Web formats based on XML allow for semantic annotations like the speaker of each speech in order to capture this debate structure, as well as the related content of debate. Up to now, unleashing these powerful cues required mastering a complex querying language, and significant “programming” skills, yet modern insights in exploratory search on structured data hold the promise to bring this power into the hands of researchers and the general public. The resulting tools exploit the semantic annotations to bring out what remains hidden in the plain text: the actual political process and strategies within a debate, as well as how the politics evolved over time — of politicians, parties, and the political system as a whole — providing a new data-driven perspective on our political history. The ExPoSe brings together two related research lines that have attracted considerable interest in recent years. On the one hand, the Dutch parliamentary proceedings have been harvested and semantically enriched in the Political MashUp project. On the other hand, the README project has led to new insights in searching structured collections (in particular those in archives museums, archives, and libraries) by both experts and novices, which suggest novel exploratory search methods that are tailored to their specific demands.
    Jaap Kamps (associate professor of Information Retrieval, UvA), Maarten Marx (Informatics Institute, UvA) and Spinque (private partner).
  • Digital Canal [proposal and presentation]
    In the seventeenth century Amsterdam stood out as the predominant center of cultural production in the western world. Paintings, books, prints and maps were produced on a scale that was unparalleled elsewhere or before. Recent and current research on this subject has a specific focus on the spatial dynamics behind the clustering of these industries. Theories from the field of economic geography offer a variety of explanations for the spatial concentration of economic (and cultural) activities, like ‘economies of agglomeration’, ‘knowledge spillovers’, and so forth. These effects can be observed and investigated on different geographical scales, ranging from global scale to street level. This project aims at developing tools for a better mapping and understanding of the spatial clustering of these industries within the city. Over the past two centuries art and book historians have gathered a wealth of data on the where abouts of painters, printers, publishers, engravers and book and art collectors in seventeenth century Amsterdam. And due to continuous research in the Amsterdam City Archives, new addresses are added to this stock almost every day. However, these data are not particularly fit for spatial analysis. Most of the available addresses take the form of textual descriptions instead of zip codes and house numbers, let alone geographical coordinates. Hence we have to start our spatial analysis with expressions like ‘op de Dam in de Wackere Hondt’ and ‘aen de Singel bij de Beulingstraet’. Geocoding is the process of translating textual descriptions into geographical coordinates. Existing geocoders (e.g. the one used by Google Maps) work pretty well with standardized toponyms, zip codes, house numbers, and so forth. However, when we deal with seventeenth-century addresses we are confronted with variation in spelling, synonymy, homonymy and changes in the location and nature of the referent. This project aims at developing a geocoder that works with non standardized descriptions and that takes into account the changing cityscape. The geocoder will be made available within a web based platform for the collection, display and mashing up of spatial data on Amsterdam in its Golden Age: The Digital Canal Ring (Digitale Grachtengordel). The data exists of addresses of painters, printers, publishers, etceteras in seventeenth-century Amsterdam, addresses of their customers, georeferenced old maps, georeferenced toponyms and historical Points of Interest, applying technologies as Sphinx search engine, OpenLayers and Leaflet, PostgreSQL and PostGIS. Marloes Scholtens (UvA), Harm Nijboer (UvA), Marten Jan Bok (UvA), Richard Duivenvoorde (Webmapper), and Edward Mac Gillavry (Webmapper).

VU projects

  • First Aid Online [presentation and report] First Aid Online investigates the effectiveness of an e-learning environment (mijnehbo.nl). In the first part of the project the effectiveness of a traditional offline EHBO-course is compared to the e-learning environment. In the second part of the project the effectiveness of different information presentations (animation with written and spoken text) in the online learning environment is investigated.
    Vera Mast (embedded researcher), Charlotte van Hooijdonk (supervisor, Language and Communication VU), and Oranje Kruis (private partner).
  • Creative Industries in the Dutch Republic [proposal, presentation and report&appendix] Creative Industries in the Dutch Republic is integrating different datasets on books on the Dutch Republic in 1660-1750 in order to research innovative market strategies of the early modern history of the creative industries.
    Fernie Maas (embedded researcher), Inger Leemans (supervisor, History VU), Albert Merono, Wouter Beek and Rinke Hoekstra (Network Institute, VU), and Amsterdam University Press (private partner).
  • BiographyNED [presentation] BiographyNED provides an analytical tool to connect historical characters and events within the Biography Portal of the Netherlands (Huygens/ING), thus building a semantic layer on top of the current portal in order to enrich our sources and analytical tools for history writing. This is a project from the Netherlands eScience Center and is not a public-private project as such.
    Netherlands eScience Center and Susan Legêne (History VU).
  • Archiving Interactives [report and policy advice in Dutch] Archiving Interactives focuses on issues concerning the cataloging, storage and accessibility of interactive artistic productions which are a growing sector within the Dutch e-culture. What do these new forms of interactive media art as heritage demand? And more specifically: How can the process of production and participation be captured that characterizes the new products of e-culture? To answer this question, Archiving Interactives aims to generate insights into the specific mediality and, develop analytical tools for unlocking, preserving and making accessible these interactive productions.
    Jesse de Vos (embedded researcher), Martina Roepke (supervisor, VU) and Beeld en Geluid (private partner).
  • Tailor-made News
    The audience’s relationship with news is shifting: news users increasingly resort to mobile devices such as smartphones, tablets, and laptops, and social media like Facebook and Twitter for their daily news fix. On these mobile and social platforms, news users also more and more have the possibility to act as their own gatekeeper, selecting only the news they are interested in. Consequently, they not only have greater control over the time and place of their news use, but also over the content and presentation of the news. Although news organizations feel pressure to keep up with the latest developments in media technologies, they have little knowledge of the news preferences of their audiences on new mobile and social platforms. Do users want the same news selection when using their iPad as when watching television? Which news stories or news genres do they want in their Facebook ‘feed’ and in what form? To what extent do they want to customize the news app on their smartphone? Taking as a focal point the perspective of the user, this study seeks to provide news organizations with insights on how to more effectively meet the changing and varying demands of their audiences for different news devices and platforms. It triangulates the results of an analysis of 20-25 in-depth interviews with news users with 5 interviews with (chief)editors and policy makers of the Netherlands’ biggest, public newscaster NOS News with an inventory of new news products on mobile and social media.
    Tim Groot Kormelink (researcher) and Irene Costera Meijer (supervisor, VU).
  • Transparency in Journalism
    Transparency is a relative new professional norm in journalism, having gained extra relevance since the digitalization of news and the rise of social media. Transparency has been suggested the ‘new objectivity’ in journalism. This paper analyses the functioning of different manifestations, instruments and practices of transparency — understood as the disclosure of organizational procedures and journalistic decisions, ranging from one-way openness to interactive, participatory communication with news users. In journalism, it is widely assumed that the public and democracy will benefit from transparency. The absence of studies that directly address the experience and evaluation of transparency is therefore remarkable. This paper addresses the discrepancy between the assumption of transparency as a criterion of journalistic quality and the lack of ‘proof’ for it. It analyses the value of transparency from the perspective of its users and its main producers, the news organization. The paper is based on a case study which focuses primarily on the conditions for and manifestations of valuable transparency, and the credibility and public responsibility of the Dutch national public newscaster: NOS news. How may news become more valuable for its users when news organizations become more transparent in the digital age? Qualitative audience research (in-depth interviews with 20-25 news users) is triangulated with 5 interviews with professional journalists and NOS-policymakers, and an inventory of the current (national and international) transparency manifestations. This methodological triangulation will result in patterns and interpretative repertoires, with which conclusions can be drawn about the use and possibilities of transparency for (public) media organizations.
    Martje Doeve (researcher) and Irene Costera Meijer (supervisor, VU).

KNAW projects

  • Dynamic Drawings [proposal and presentation] How can technical operations and workings of scientific instruments described in historical scientific texts in printed publications via the web in a user-friendly way be visualized and permanently archived for (re-) use in enhanced publications? Dynamic Drawings visualizes old drawings into colorful 2D or 3D animations and explores new ways of publishing in which publications in print and digital form are used complementary.
    Charles van den Heuvel and Maarten Hoogerwerf (researchers KNAW), Tiemen Cocquyt (researcher Museum Boerhaave), Wild Card (private partner), and Brill Academic Publishers (private partner).
  • TINPOT [proposal and presentation] TINPOT develops TweetGenie, an algorithm that is capable of predicting age and gender of Twitter users based on language used in their accounts. Using this while following product rumors from beginning to end, a better analysis can be made of the narrative conduits and the people involved.
    Theo Meder (researcher Meertens Institute) and Teezir (private partner).
  • Radical Political Representation [proposal and presentation] To what extent can we apply modern information technology to digital, visual (non-photographic) source material (in this case political caricatures) for doing historical research? Radical Political Representation uses crowd sourcing to annotate cartoons from the national socialist newspaper Volk en Vaderland (1933-1945).
    Kees Ribbens (researcher NIOD Institute), Junte Zhang (researcher Meertens Institute), Peter Romijn (researcher NIOD Institute) and Zeezeilen (private partner).