DiscoverText

DiscoverText is a cloud-based, full-featured text analytics and data science platform designed to allow users to easily accumulate, sanitize, and analyze large volumes of unstructured text data from just about anywhere. Its intuitive point-and-click graphical user interface allows users from researchers and academics to legal departments and market analysts to sort and analyze free text and associated metadata from customer feedback, surveys, email, chat, public comments, and social media feeds like Twitter and RSS feeds. The software contains dozens of multilingual text mining and machine learning features that offer simple and advanced workflows for text classification, sentiment analysis, and topic modeling. Through seamless integration with common data sources and support for flexible importing, DiscoverText streamlines the otherwise arduous process of preparing and cleaning noisy text data for meaningful analysis. One of the core strengths of DiscoverText lies in its hybrid data science model, wherein human annotation and machine learning work together to deliver high-quality, interpretable outputs. The users can build reusable custom machine classifiers, or "sifters," to be able to readily identify and categorize useful items in huge datasets with far less time devoted to initial review and categorization. The platform's workflow involves crowdsourcing, iterative measurement, adjudication, and annotator ranking to generate gold-standard training sets for machine learning. Its CoderRank patent assigns a score and ranking to human annotators based on time to ensure that both human and machine contributions are consistent and reliable. This new adjudication process is particularly valuable in research at universities, where transparency and reproducibility are of highest importance, and for legal staff who require defensible, auditable processes for eDiscovery and compliance. DiscoverText also contains advanced eDiscovery and information retrieval features, making it especially suitable for legal and regulatory environments. The software features automated clustering and deduplication of near-duplicate documents, which allow users to quickly visualize the landscape of large sets of data, whether identifying viral tweets, form letters in public comment streams, or duplicate opinions in surveys. Its interactive machine classifier histograms and purposive sampling techniques allow data science teams to focus human coding efforts on the most valuable items, further accelerating the training of machine classifiers. Its other features, such as document redaction, Bates-stamped PDF output, and spreadsheet-indexed sets, ensure that sensitive information is processed securely and easily. Utilized by researchers and attorneys, DiscoverText shortens analysis timelines from weeks or even months to just hours, making it the preeminent solution for open, collaborative, and scalable text analytics.

DiscoverText stands out from the rest through its particular mix of state-of-the-art machine learning, human-annotated collaboration, and interactive data science tools, all offered in a user-friendly, cloud-based system. Perhaps its strongest feature is the text analytics hybrid method, allowing customers to build reusable, proprietary machine classifiers—"sifters"—that rapidly identify and classify relevant items from enormous, unstructured datasets. This hybrid solution combines both crowdsourcing and automated machine learning to enable iterative measurement, adjudication, and ranking of annotators to create gold-standard training sets. Its patented system, CoderRank, inspired by Google's PageRank, ensures that human annotation is not only accurate but continuously improves machine learning performance over time, a promise rarely kept by other text analytics vendors. The other key differentiator is DiscoverText text mining, data cleaning, and eDiscovery. The platform can ingest data from a huge range of sources—social media, surveys, emails, chat, and public posts—and has extremely effective tools for deduplication, clustering, and intelligent search. These features allow users to quickly map the landscape of large datasets, identify trending social media tweets, and cluster similar near-duplicate opinions or templated letters for speedy review. DiscoverText's point-and-click interface is also designed to be within the grasp of users with varying technical aptitude, thus both academic researchers as well as legal teams can carry out sophisticated analyses without the requirement of extensive training. Annotative, redacting, and Bates-stamping PDF collection creation features provide additional value for legal and regulatory processes. Transparency and collaborative features of DiscoverText also place it among the best for research and legal purposes. DiscoverText supports real-time collaboration, project management, and adjudication, allowing several users to contribute, review, and validate coding schemes. Its process not only accelerates machine classifier learning but also renders results reproducible and defendable, essential in academic papers and court cases. The platform has been known to shorten analysis times from weeks or months to hours, both rapid and precise. With open access and direct support for students and faculty, in addition to high customer service and ease-of-use ratings, DiscoverText remains a clear best-of-class choice for text, metadata, and social network analysis, offering the flexibility, reliability, and transparency that cause many of its competitors to seem like also-rans.

• Small Businesses (1-99)
• Mid Market (100-500)
• Enterprise (500+)