DiscoverText (discovertext.com) is a proprietary software tool created by Texifter that supports collaborative search, filtering, duplicate detection, human coding, and machine-learning. At its core, DiscoverText supports measurement techniques that allow researchers to better understand the ability of individual coders (annotators) to reliably classify text data (inter-rater reliability) as well as the accuracy (validity) of individual and aggregate human observations. The U.S. Patent and Trademark Office issued a patent (9,275,291) for Texifter’s method and system of “CoderRank” on March 1, 2016. CoderRank enables researchers to report on the accuracy of human observations and it also contributes to the development of gold standard training sets for machine-learning.
DiscoverText is a web-based, text analytics toolkit. It combines a collaborative work flow with the ability to develop and reuse custom machine‐learning text classifiers. These Sifter™ technologies help users to quickly conduct better and more accurate analysis at scale. The text analytic process reveals common themes throughout unstructured and semi‐ structured text input and it also helps identify unique, infrequent or unexpected findings.
DiscoverText is a an API-enabled social media collection tool. Users can import "live feeds" from Facebook, Twitter, and other sources, as well as large email collections for e-discovery. It is also integrated with SurveyMonkey, supporting analysis of large numbers of open-ended answers and the associated metadata.Our de-duplication and clustering procedures render the raw data into a form that is more easily understood, reported, and acted upon.
Using the premium Gnip historical PowerTrack, we can access every non-deleted Tweet in history. The DiscoverText application shows individual Tweets in the native Twitter display. This means, when looking at a Tweet in DiscoverText, a user sees any images, media link previews, and a live display of the accumulated re-tweet and favorite counts.