Voting closed 15 Dec. 2012. 77 Liked
Title: Maphub – old maps in globally connected data networks
Old maps are a record of the past, exposing features people might want to tell stories about. We built Maphub, which is a prototype Web application that enables them to do so by creating annotations on digitized high-resolution historical maps. By semantically tagging regions on the map, users create associations between their annotations and resources in open Web-based data networks. These associations are leveraged to enable multilingual search and to generate overlays of historical maps on modern mapping applications. Contributed annotations are shared on the Web following the W3C Open Annotation specification. Maphub is built on the Linked Data technology stack and a first demo has been setup with the Library of Congress’ historical map collection. The high-level goal of this project is (i) to showcase how existing Linked data can be used and integrated into existing scholarly processes, such as annotations, and (ii) vice-versa, how applications could support users in contributing data and knowledge to broader, open data and knowledge networks.
Historic maps reflect geographic information and also the attitudes, perspectives, and beliefs of different times. Their geographic accuracy tells us much about the state of geographic knowledge and technology at the time of their creation. Thousands of maps have already been converted to digital form and made available online and accessible to scholars and the broad public.
When viewing digital maps users often have stories to tell. They might know something about the context of a map, might be able to identify places or landmarks, or might have comments that could complement historic maps as records of the past. However, most current map hosting environments don’t allow users to contribute their stories.
We believe that allowing annotations on historic maps is a possible way of recording these stories. If we also connect named entities mentioned in these stories with other related resources in open globally connected data networks, such DBpedia or GeoNames, we can capture the context of these stories and, through named entity disambiguation, we can connect the stories of different individuals. This user-contributed information, in combination with curated metadata, is a valuable source for search and retrieval or any other data analysis task.
The current Maphub prototype is the result of a demonstration experiment carried out as part of the Open Annotation Collaboration. It showcases how users can annotate historic maps and connect these annotations with resources in open data networks via a function, called semantic tagging. Tagging creates relationships between annotations and Web resources. These relationships are leveraged to supplement user-contributed annotations with additional information from the Web, enabling functions such as multilingual search and retrieval over historic maps. Collected annotation data are contributed back to global data networks by exposing them as dereferenceable Web resources following the W3C Open Annotation specification.
Using the commenting tool (“Annotate”), it is possible to draw shapes on the map, such as rectangles, lines and polygons. These shapes mark a region of interest, and once the user has finished drawing, a popup will ask them to enter their comment and tell their story about that region in a free text field.
While the user is writing text, the Maphub system analyzes the input and proposes possibly relevant semantic tags. These tags are links to Web resources such as Wikipedia articles and are suggested by querying open data sources such as Wikiminer or GeoNames.
To disambiguate the meaning of these tags, each tag carries a short description of its subject, available as a tooltip. Furthermore, we allow the user to accept or reject individual tags before annotations are being saved. In the background, Maphub then dereferences the URIs of accepted semantic tags and adds subsets of the retrieved data representations to the index. This way users can, for example, search for maps in any language supported by Wikipedia.
The georeferencing tool (“Control Point”) allows users to associate known locations on a historic map with location resources provided by the GeoNames web service. As soon as we have at least three control points, we can compute a model for translating between the x/y pixel dimension of the digitized historic map image and real-world projections used in modern mapping systems. This allows us to visualize historic map overlays in Google Maps and project maps onto a three-dimensional globe, provided by Google Earth.
Sharing collected annotation data in an interoperable way was another major goal of this demonstration experiment. Maphub is an early adopter of the Open Annotation model, which is currently specified in the W3C Open Annotation working group. It demonstrates how to apply that model in the context of digitized historic maps. As described in the Maphub API documentation, each annotation becomes a first class Web resource that is dereferencable by its URI and therefore easily accessible by any Web client. In that way, while users are annotating maps, Maphub not only consumes data from global data networks – it also contributes data back.