Taking «Auto-Tagging» to another level: The successful integration of iMatrics in Livingdocs.
Let’s face it: in today's fast-paced, advanced and digitalised world we can hardly expect anyone to keep up with current demands by simply analysing and tagging texts manually. And no longer will anyone argue the importance of valuable metadata for media companies. Metadata has become almost as important as content itself. The online world is meanwhile controlled by algorithms to such an extent, that an article without metadata can result in being of very little value and in that sense become close to invisible in the web.
Publishers require reliable and fast solutions in order to cope with the ever increasing standards, when it comes to analysing, structuring and tagging. They need reliable auto-tagging systems that can automatically tag text-based assets and learn intelligently with the most current Machine Learning and AI methods. These are systems that allow them to use and generate tags in real-time without anyone manually applying them, but will at the same time also permit and integrate easy and fast manual editing and adjusting.
We are proud to announce that this mission is completed!
For now, of course ;-) ...in that very obvious sense of setting a solid foundation for many intriguing future developments:
Livingdocs has successfully integrated the highly advanced auto-tagging system by iMatrics.
«The Livingdocs team is amazing! It is inspiring to see how they embody their vision for the media industry in the products they develop and we are thrilled to contribute to this environment with our analytical tools. It has been a true joy to work with them and we are excited for what’s to come.»
iMatrics is a Swedish software company based in Linköping. It provides solutions and services for advanced text and data analysis. Its core principle is that companies should be able to effectively use all its stored information. In order to help companies achieve this goal, it provides services such as content identification, recommendations, data analysis reports. All services are tailored to its customers specific needs. iMatrics’ natural language technology automatically classifies categories, topics, subjects, organisations, people, and places in news content, which enables fast and actionable insights for customers. Its language independent text analytics & AI technology has found especially high recognition in the media industry.
This November, CH Media, one of the leading Swiss media companies, is our first customer to go live with iMatrics’ solution integrated in Livingdocs. Very soon, in a future post, we shall gladly inform you in more detail about their specific and impressive integration.
Read on to learn more about how exactly the integration of iMatrics’ system works.
In Livingdocs, every article is a structured document. It contains the actual content as well as a variety of metadata. The kind of metadata that is applied varies between different systems as different setups and publications have different needs. In iMatrics the recognised categories, topics and entities are called concepts. These concepts are stored in Livingdocs in one of these metadata properties per document, even per version of a document. That allows for implementation of various specific use cases on top of it.
To analyse the content, all the textual content gets extracted from a document while it is edited by one or multiple editors concurrently and is sent to the API provided by iMatrics. The natural language processing system at iMatrics then returns a set of categories, topics and various entities. In Livingdocs, these results can be seen and manually fine-tuned by editors in case certain topics are unfitting or a specific one is missing. Should a specific topic be missing and not yet be defined in iMatrics’ system, it can get suggested to be approved or declined by an iMatrics administrator at your newspaper. Since the News Media broadly mentions topics long before they become part of our collective history, this will probably happen regularly.
To keep things simple for most users, that’s everything regular users in your Newsroom will ever see of iMatrics.
Since iMatrics provides a unique corpus of concepts developed in close collaboration with every customer we built a dashboard to administer the concepts known to iMatrics right into Livingdocs in close collaboration with NZZ Technology. This way, expert users in your company can edit this corpus as soon as any need emerges. Don’t worry, the corpus is mostly managed by the team at iMatrics, but it’s still easier to just add something that is missing yourself instead of having to call your contact at a service provider.
If you would like to enable these features in your Livingdocs instance as well, get in touch with the fantastic people at iMatrics via email@example.com. They will be more than happy to give you a demo for you to see and understand how it works.
If you have ordered the service from iMatrics, you can enter the necessary credentials in the Livingdocs project setup screen, add a metadata property to the content-types you want to get automatic tags for and you are ready to go. Every future document will automatically get tags describing your content.
The use cases for the iMatrics integration vary from automatically generated topic pages, to automated or semi-automated next-read suggestions to a deep integration into your content planning to support you in covering all important topics. Why not have a tag-cloud with the most covered topics of last month shown on a screen in your cafeteria?
At Livingdocs, we will continue to deepen our partnership with iMatrics to serve you with great insights into your documents. Together we structure all your content so you can maximise its value for your customers.
For more information on iMatrics please visit imatrics.com