1. articles published between June and December 2012 by major news outlets were scraped from the web.
2. A dictionary of all named entitties was produced to later recognize names of people and companies from text.
3. An associative network was produced in this fashion: each article containing the word "corruption" and all its forms was processed paragraph by paragraph. A list of named entities was extracted per paragraph. These entities were all connected in a subgraph. Then the subgraph was integrated into master graph, keeeping date information intact.
4. The master graph,containing information (all entities and relationships) was visualized as a dynamic network in Gephi and animated.