What is Connected Data as a concept? Who is interested in Connected Data? What problems does Connected Data solve? What skills are used in Connected Data?
Connected Data as of July 2017 has been running for over a year with very successful conference and 9 meetups held to date on a range of topics. These have included Knowledge Representation, Semantics, Linked Data, Graph Databases, Ontology development and use cases and industry verticals including fraud analysis, recommendations, telecoms and finance. Yet the group has never had a particularly formal terms of reference or description defining what Connected Data actually means. Some would say this is something of an irony for a group so focused on semantics, schemas, definitions & structure!
At the July meetup group organiser James Phare made an attempt (with some humour and something of data journey thrown in) to achieve something resembling a definition and terms of reference for the group. Whilst it’s clear that caring about the connections in data has always been a common trait amongst group members the methods for modelling data, the approaches and technologies used vary. Often this leads to great opportunities for knowledge sharing and heathy debates for instance about for instance schemaless approaches vs schema centric approaches.
The objective of the ‘experiment’ was to create artefacts to help people understand what Connected Data is and what the minimum common features are that group members possess. James started with pulling together a list of terms for the technologies, practices, standards, roles and other concepts Connected Data as a group embodies. A good old word cloud was then used to try and ‘visualise’ this content. What resulted can only be described as a colourful mess. Any useful information regarding relationships, any context was missing. As a next step James showed how he started to organise these concepts and words into groups using a very loose taxonomy and some rules. Again this began to yield a bit more insight as to what the focus of the group was but still lacked detail around exactly how these concepts were related to each other. Hierarchical relationships became the next obvious point for aiding understanding by adding wider and narrower links between concepts, however again there were interesting things missing such as related topics and links between different entity types such as which roles typically use which technologies. Eventually a network or graph began to emerge – perhaps this was the single binding characteristic that all group members shared?
However soon another dimension of the problem emerged – that of use cases or requirements. To what degree are these common characteristics or beliefs dependent on the projects that people are working on? For example whilst a schemaless graph might be an excellent choice for building a recommendation engine for a narrow product domain perhaps more precise semantics and rules are required for more complex requirements such as Knowledge Management? Perhaps as a group we all have more in common than we previously thought?
Either way eventually with the aim of succinctness in mind James proposed the following as single term to describe what defines Connected Data and the people that operate within this space. We hope you like it but as an open community we are always open to suggestions and enhancements!
“Connected Data is a collection of related Disciplines & Technologies where the relationships, rules, structure, meaning & context are treated as first class citizens”.