Group items tagged

 DiscoverText's latest feature additions can be easily trained to  perform customized mood, sentiment and topic classification. Any custom  classification scheme or topic model can be created and implemented by  the user. Once a classification scheme is created, you can then use  advanced, threshold-sensitive filters to look at just the documents you  want.  You can also generate interactive, custom, salient word clouds using  the "Cloud Explorer" and drill into the most frequently occurring terms  or use advanced search and filters to create "buckets" of text.  The system makes it possible to capture, share and crowd source text  data analysis in novel ways. For example, you can collect text content  off Facebook, Twitter & YouTube, as well as other social media or RSS  feeds.  Dataset owners can assign their "peers" to coding tasks. It is simple  to measure the reliability of two or more coder's choices. A  distinctive feature is the ability to adjudicate coder choices for  training purposes or to report validity by code, coder or project.

Short essays on various topics, posted links to Twitter.

Boredom for a subject does not reflect a defect in the subject, but in our understanding of it. In the ears of the ignorant, a foreign language is a monotonous barrage of meaningless intonations, but knowledge of its grammar transforms sound into speech, capable of conveying Shakespeare's or Plato's meaning. The surface of Mars seems to me a tiresome landscape of red dirt, but to an astrophysicist who speaks the obscure language of rocks, it is a crossword puzzle written by the Big Bang. We protest to the passionate not to bore us with details, not realizing that lack of details is precisely what bores us, for details reveal the richness and inner coherence that are invisible from a distance, as a microscope reveals teeming life in a drop of muddy pond water. ...