What is Context Mining?
Swan, as the Big Data & Analytics Lab of the Bisnode Group, has designed a structured approach to R&D, which combines the creativity required at the exploration phases, and the pragmatic business applications for Bisnode customers. We call this approach Context Mining.
Context Mining merges in its very words two important aspects:
Context, which is at the core of our mission (bring Contextual Intelligence to Business Information and Consumer Marketing), and
Text Mining, and natural language processing techniques, which is the core capability allowing us to reveal context.
The fundamental exercise consists in collecting the right data and in combining it with other sources in order to breed new information. Context appears through the relationship between disparate data, and through the enrichment (we mean: new information) that comes out of it. In that sense, we mine the context of data.
Our approach deploys four main layers, each one being interlinked with the three others.
Data Sources – one of the core specialty of our team is the capacity to find the right data sources, mostly online, and to manage ways to collect them (legally, of course). We collect worldwide news, social data, open data and « closed » data (negotiated or purchased, including corporate data).
Data Science – on top of those data, we apply the most-advanced data science algorithms to extract value, create new information, reveal hidden insights. Our team is pushing the boundaries of machine learning, NLP/Text-Mining, Network Science and Data Engineering.
Data Components – the right data with the right science produces Data Components – small pieces of technology that fulfill a specific data-driven task with some magic (see some examples here). The list of Data Components is growing every month. A Data Component can be used as such to run a task, or embedded in a product as a core engine.
Products – the ultimate phase of the Context Mining approach is the development of sellable and scalable products that go through the Bisnode Group’s product management channels to the end customers. Our first release is Bisnode Business Events.
The Context Mining approach shapes the way we are organized, the way we work, progress and report. It is also the lighthouse that we, data sailors, never loose sight of – if our work does not bring context, then it is not worth it, and we move on.