Decoupling Drupal - Data Engineering To Make Sense Of Unstructured Data Representing A Million Voices For The United Nations
The session will cover various aspects such as visualization of linked data and also discuss the road map to implement visualizations to help identify key influencers in conversations, bi-axial representation of sentiment analysis expressed in discussions vis a vis documents, trending topics, opinion mining and more.
For almost one year, the United Nations Development Group (UNDG) collected the perspectives on the 'World We Want' from over one million people around the globe, enabled by national level consultations, thematic dialogues and surveys in order to help member states shape a successor framework to the Millennium Development Goals (MDGs) beyond 2015.
The next step was to visualize and discover the output, mostly unstructured, generated by these 'million voices', most of which made available in a Drupal platform.
In the session, we will discuss how Azri deployed a decoupled solution to make sense of unstructured data across multiple data sources, implementing various NLP techniques ranging from word sense disambiguation to leveraging text processing libraries for classification and semantic reasoning, thereby enabling use of all available unstructured data in a structured fashion.