In their 1977 short film Powers of 10, Charles and Ray Eames demonstrate that different kinds of understanding are possible at different orders of magnitude. When designing data-driven systems, it is crucial to analyse data at the human scale as well as at the mass aggregate scale.
Today we have a model for understanding data at different levels of magnitude: Google Maps. We can zoom in and out of geography, and are able to distinguish and analyse continents with a level of detail appropriate for the scale. That does not include minutiae such as streets, houses and parks. In that frame, we can identify data that is grouped at the level of Europe, Africa and so on – and we do not require more complex insights at that level. We zoom in to distinguish features such as cities and their geospatial relationships. We are able to orient ourselves with more and more levels of granularity, or we can zoom out again to get a sense of the overall picture.
At the human scale, there is a very different requirement of data mapping than there is at a global scale. As such, the notion of ‘zooming in’ provides a very good metaphor for how we should design intelligent data systems.
Data is not only stored in the cloud, it is also analysed in the cloud. Smart IoT systems use Big Data filtering, create ontologies and classifications in order to make sense of that data, consider the context of data usage and use AI to train systems to recognise patterns in data. The wealth of aggregate data accumulated by the Allternet provides, in itself, opportunities for understanding at a high level of analysis, but that analysis may not be relevant at the human scale.
In order for ‘Internet of Things’ projects to be validated, it is essential to run pilots deploying agent-driven applications. In this way, it will be possible to test, for instance, a ‘System of Systems’ in physical space, in relation to a scale comprehensible and useful to the people using the devices within that system. In this way, these projects and systems are contextualised and understood within the broader Allternet space.
There are those who advocate creating a system of systems in abstraction. There is another school of thought that believes we should start from the users. Neither of those two is better nor more important than the other. Instead, it’s about the rules or assumptions that can be made at each level. At the level of reconnaissance, there is a more abstract relationship with data which is about describing contours. At the level of the individual, the relationship is with the person and their specific needs and requirements. Not only are both valid, but there are multiple layers of understanding that can be reached at different levels of magnification.
It’s important not to make assumptions about somebody sitting on the street based on data that is mapped from the perspective of an altitude of 10,000 feet. When you’re sitting next to that person, you will have a very different understanding of what data is useful to you.
Designing data-driven systems is about creating truly intelligent systems that understand and appropriately respond to scale (as with the Powers of 10) – as well as to time, since data takes on different kinds of meaning over time. If you create one set of descriptors at a particular time, you will inevitably need to renew those descriptors when conditions change. The Allternet is, in this respect, like a living ecosystem.
As with data – so with ideas. An idea is always a result of particular affordances and parameters that are on offer at that particular point in time. In the case of EU-funded projects, this is usually mapped up front, rather than allowed to evolve. Any intelligences that we can draw from these projects change too. Because EU projects are locked in the first moment, they struggle to create a good business model. The project is cemented in the past before it has begun. Good business is always a living ecosystem. It needs to continually innovate in order to survive, keep ahead of competition, and reinvent itself.
Understanding data at scale (and over time) reflects the fact that the Allternet acts as a living ecosystem. From that adaptive, reactive and context-aware starting point, novel and disruptive IoT business models can be supported.