meaningasaserviceDaedalus (which The Semantic Web Blog originally covered here) has just made its Textalytics meaning-as-a-service APIs available for Excel and GATE (General Architecture for Text Engineering), a JAVA suite of tools used for natural language processing tasks, including information extraction in many languages. Connecting its semantic analysis tools with these systems is one step in a larger plan to extend its integration capabilities with more API plug-ins.

“For us, integration options are a way to lower barriers to adoption and to foster the development of an ecosystem around Textalytics,” says Antonio Matarranz, who leads marketing and sales for Daedalus. The three main ecosystem scenarios, he says, include personal productivity tools, of which the Excel add-in is an example, and NLP environments, of which GATE is an example. “But UIMA (Unstructured Information Management Applications) is also a target,” he says. The list also is slated to include content management systems and search engines, among them open source systems like WordPress, Drupal, and Elasticsearch.

The new Excel plug-in lets users analyze tweets, posts, opinions in forums, surveys, news, and so on, for sentiment, entity detection or text classification from right within their spreadsheet “This is a resource that lowers the semantic technology adoption barrier for business users, although developers can also benefit from it — for quick evaluation purposes, for example,” Matarranz says,   We’ve seen a demand for this kind of tool.

He notes that market research agencies are a good example of users that can get value out of the plugin.  “Many agencies are not very technology–oriented and they more and more need to extract insights from unstructured content: free-form comments in surveys, social media conversations,” he says. “They need to analyze for their customers, say, 10,000 tweets or comments per month and classify, extract mentions, sentiment… The Textalytics add-in for Excel enables to do just that in a very convenient way, integrated with their typical data analysis tool.”

The plug-in for GATE is mainly targeted at the high-end of the market, both in academic and business scenarios. As examples, Matarranz points to the BBC tagging contents in its publishing process using GATE and to pharma companies, such as Astra Zeneca or Elo Lilly, that have used the platform to annotate MedLine abstracts. Advanced users, he says, apply GATE to design, prototype, share and evaluate text processing pipelines that incorporate various software components.

“For Textalytics users, GATE helps them escalate and make more agile their development processes. For GATE users, Textalytics provides advanced, verticalized, multilingual semantic analysis (including high-quality coverage of less-common languages like Spanish),” he says.

Plans also are underway to offer new releases for its APIs for Media Analysis and Semantic Publishing. The Media Analysis API provides a series of high-level services (such as detecting buying signals and extracting customer insights) that can be implemented in all media and publishing business activities. In its latest version it will include functionality for trend detection (specially the early detection and characterization of previously unknown and unexpected conversation themes), Matarranz explains. Also in the upgrade is advanced profiling of users, such as demographic/psychographic characterization based on their public profile description, connection network and conversation themes.

The new Semantic Publishing API, he says, will include functionality for dynamic content generation, relation and personalization, “so that publishers can automatically enrich, grow and add value to their contents, thus increasing user engagement and content discoverability.”

Another new high-level API is scheduled to be published as well. It will be devoted to Voice of the Customer (VoC) and Customer Insights. “The idea here is understand customer feedback (especially of an unsolicited, unstructured nature) from all kind of channels (contact center interaction, surveys, social media),” he says. “We plan to extract high-value insights such us brand perception, customer journey, buying/churn signals, corporate reputation indicators… We have several of these features in beta with some real customers.”