In our continuing focus on Analytics APIs that help enterprises come closer to the customer by easier integration through APIs, here is an interview with José Luis Martinez of Textalytics. Textalytics is a brand by Daedalus, a Spanish SME that has been working in the field of natural language processing since 1998.
ProgrammableWeb (PW) -What does the Textalytics API do?
José Luis Martinez (JL)- Textalytics provides meaning as a service. Textalytics delivers meaning extraction from multilingual texts, allowing an easy integration in any application. We say ‘easy integration’ because we offer vertical APIs, specifically developed for an application domain where users do not have deep linguistic knowledge (for example our Semantic Publishing API) and horizontal ones, for those developers who want a complete control of the linguistic processing. On the other hand, when we say ‘meaning extraction’ we do not only include the identification of language elements such as concepts, named entities, relations among them, dates, addresses and many more structures, but also the understanding of the content, saying if there is an opinion in the text, if it is good or bad or ironic …
You can think about a lot of applications where meaning extraction can contribute with really useful functionality. Consider these situations: your customer writes you an email complaining about your product, a reader of your blog writes a comment to one of your posts, or a citizen calls an emergency phone to warn about a fire in a building. What similarities can you find in these scenarios? All of them need human processing at the other side of the channel. Textalytics provides ways to better automatize the way all these interactions are processed by automatically understanding what the user or customer is saying.
This functionality is offered in a SaaS mode through a set of APIs.
Textalytics Media Analysis is the first vertical API we offer. It analyzes mentions, topics, opinions and facts from any type of media. The goal is to make the semantic analysis of blogs, forums, microblogging, news, etc easy. Results of this semantic analysis can help in the identification of customer insights, buying signals or corporative reputation facts.
Horizontal products are grouped into Textalytics Core API. This set of APIs is formed by linguistic processing components or building blocks for broader applications. It includes semantic analysis, text classification, topics extraction, speech recognition, proofreading and many more. Any developer can easily integrate these functions into its applications.
PW-What are your plans for the forthcoming Semantic Publishing API?
JL- Textalytics Semantic Publishing API will offer (online and paper) publishers semantic tagging tools to enrich the contents they generate and, at the same time, to increase content reuse in their processes. The world is changing quickly in the publishing market, so any help taking the majority of the content generated by publishers can have a dramatic effect. New ways to reuse and relate information in this environment are needed. On the other hand, thanks to the Internet, new individual or small publishers, devoted specifically to online markets, are appearing. With their limited capacity to cover all steps in the publishing process, tools helping them to automatically enrich and relate their contents can make a difference. One step ahead, another critical success factor in this market is the Web 2.0 approach. Users want to talk, visitors are willing to leave their thoughts about what they are reading but, at least in Europe, the online publisher is responsible of the comments written by their readers. This is the reason for Textalytics to offer automatic forum moderation, to assure that disrespectful user comments do not get online. This functionality will also be part of Textalytics Semantic Publishing. According to our roadmap it will be ready by the beginning of December.
The next product in our roadmap will be the Voice of Customer API, in march. This API will allow customer monitoring in several ways: buying signals identification, spotting the phase of a user according to the customer journey approach; contents according to CRM needs; content moderation for user given comments and opinions; and corporate reputation measurement according to predefined reputation models.
Three months later we plan to publish a FAQs API to facilitate frequent asked questions management through natural language for any website.
PW- What kind of languages and platforms does your API support (i.e Python, R, Ruby etc)?
JL- Textalytics follows REST design principles, so it is really easy to make calls to any of the services from any programming language and development environment. To facilitate integration to developers, Textalytics also makes SDKs available for Java, PHP and Python. Of course, these will not be the only SDKs to be deployed; in future weeks integration with .NET and Ruby will be published.
PW- How did you come across the basic idea for the Textalytics API. Describe your own and your team’s background in text analytics:
JL- Textalytics is a brand by Daedalus, a Spanish SME that has been working in the field of natural language processing since 1998. In fact, Daedalus was born as a spin-off from the Universidad Politécnica de Madrid, as a joint effort of two research groups: the Intelligent Systems Group, working in the field of artificial intelligence and natural language processing, and the Computational Linguistics Lab from the Universidad Autónoma de Madrid. With several Phds, 100% engineers, and computational linguistics in the staff, we have a deep involvement in R&D and innovation activities, always improving our resources and algorithms, testing and developing new technologies and searching for problems to solve.
Among its customers, Daedalus counts the biggest publishing groups, from Unidad Editorial to Vocento, Grupo PRISA, and some other online publishers such as lainformacion.com.
So, Textalytics is just another way to make our already available core technology reachable to the community of developers. The idea is to give everyone the ability to integrate semantic functionality in their apps with a minimum cost.
PW- Any plans for targeting Hispanic markets in North America or beyond?
JL- Sure. It is the first step we want to take in the North American market. Although our tools can process texts in English, French, Italian, Portuguese and Catalan, apart from Spanish, markets where the Spanish language has a main role are our natural ones. In a first phase we have focused on North America because we think that SaaS and text analytics technologies are more widespread there than in other regions. In a second phase we will also build some relations with South and Central American partners.
PW- How are you trying to make your API popular for usage with developers?
JL- We are completely aware that developers play a central role in our Textalytics offer so we are working hard to let them know about the capabilities they can easily integrate in their solutions. We always think of these developers when including new functionalities and defining the corresponding structures needed for the API. Among the activities carried out in the last days, César de Pablo is giving a talk in the Big Data Spain conference and we have recently organized a hackathon in the Codemotion conference. Besides, we are working to be present in the forthcoming ‘Sentiment Analysis and Human Analytics Symposium’, showing how Textalytics is helping our customers to deal with unstructured data. We plan to continue in these communication and hackathon activities to create an awareness with developers.
PW- How do partners like Mashape help with creating an API business?
JL- When we found Mashape, we quickly saw it was the best distribution channel we could try to put our first foot in the North American market. Of course, when you talk about Internet tools, geographical references lose sense but, ok, Mashape is located in San Francisco and that defines their natural environment. Mashape provided an easy way to integrate our already existing APIs into their platform; it only took us two or three days to document, integrate and test all our APIs and there were 11 of them! Now we will work together with Mashape to spread the word among developers in order to get them involved. We are really excited about that!
José Luis Martínez short bio:
Making use of the majority of their own data and web information is a must for companies today. Such data and information can take very different forms but the current technology can deal with it, regardless of the form (text, video, image or audio) or the type of system it is stored into.
Helping those companies to gather, retrieve, analyse and, ultimately, manage all the data they need for their daily work is the Daedalus’ mission. José Luis is one of the founding partners of Daedalus.