“Media Mapping Research Portal”


Goal:


The platform will help researchers — reporters, shareholder activists, investors, and NGO’s; — find information about social and environmental issues that arise with extractive industry multinational corporations that do business in Africa. In addition, a feature will be implemented that allows researchers to connect sources and receivers of information in order to visualize and analyze the flow of information between various sources, for instance between journalists, local/international NGOs and shareholders of extractive industry multinationals. Studying these patterns will help identify gaps and help find ways to make media from a large variety of sources more accessible. Create Information Flow pattern using a variety of data points.



Proposed Workflow


  • Either directly access (geotagged)archives of news media using feeds or batch process archives of newsmedia( Use Ushahidi for event categorization and mapping, Swift Riverfor verification.
    • Build OWL ontology (then use SPARQLfor querying?)
      • On the fly — definition of sources/receivers of information and topic categories (e.g. company names, NGOs, media organizations)
      • Make standardized Bibliographic/other ontologies available as appropriate
      • Allow for both RSS or other automated data feeds, raw text and manually entered or imported metadata(COinS / Mendeley / Zotero feed
  • Mine and tag data with tags set up with keyword queries, in this case for example “Newmont” “Ghana” “Resolution” and “Shareholder” (perhaps using Open Calais for additional tags)
  • Map geotagged newsmedia and animate over time, coloring visualizations of articles by keywords, visualizing "flows" of information by using the Z axis. Using 3D to prevent edge crossings. Imagine a combination of MIT’s “NY talk exchange” and Pavel Risenberg’s “Nooblast”(uses Processing).

Questions

  • Most streamlined workflow?
  • How to verify sources?
  • How to make 3D navigation work: give user a pilot/joystick?
  • Which standardized ontologies should be used? How can data sources be reported and how can (interactive)visualizations be captured/embedded/documented when using the tool for producing reports?