DiscoverText vs KH Coder

DiscoverText is a cloud service for mining, annotating and analyzing multilingual text collections with machine learning. It can work with free text and metadata from surveys, customer feedback, CRM, chats, email, public comments, X/Twitter and RSS feeds. Users can build reusable classifiers called sifters to find relevant material and sort it into topics, sentiment or other categories. Its described process combines machine classification with human annotation, crowdsourcing, adjudication, measurement and iteration. For eDiscovery, it offers deduplication, near-duplicate clustering and document redaction that can produce Bates-stamped, spreadsheet-indexed PDF collections. Students and professors may receive free access, training and project support from the founder; academic group licenses are also listed as free. API access is limited to licensed Enterprise users granted an API key. Documented API operations include projects, archives, buckets, datasets, deduplication, clustering and tag clouds. The maker says it no longer collects Twitter data in real time and is not connected to Facebook’s open API. Professional costs $99 a month; Enterprise costs $3,000 a month.

KH Coder is software for quantitative content analysis, text mining, and computational linguistics. It can be used with free-response survey answers, interview records, newspaper articles, and other text. The analysis options include word-frequency lists, keyword searches, KWIC, cluster analysis, multidimensional scaling, co-occurrence networks, correspondence analysis, self-organizing maps, and coding cross-tabulations. Visualizations include network views with betweenness centrality and community detection, correspondence-analysis plots, heatmaps, line charts, and bubble plots. Statistical analyses can be viewed with the source-text context, and analysis can run without an internet connection. The English site lists text support for Catalan, Simplified Chinese, Dutch, English, French, German, Italian, Japanese, Korean, Portuguese, Russian, Slovenian, and Spanish. KH Coder runs on Linux, macOS, and Windows, with on-premise deployment. Users can customize it by editing R and MySQL code. The free Windows Starting Edition limits analyzable data and some features and is an older beta; the paid Base Edition is presented as the full version. The Mac Edition costs $99.00 USD once and requires Apple silicon, macOS 11 or later, and 5 GB of available disk space.