This description of a data "input" value chain assumes that data is owned by someone or by an organisation. The ISO-IEC JTC1 Study Group on Big Data has been very clear that there should be a universal attribute to data specifying its owner(s).

The data owner could be an individual: for instance, consider the case of personal data owned by a person. More broadly the data generated by objects owned by a person are likely to be owned by this person: for instance the current geographic position of my car. This means that there are expanding circles around people, with data in such circles. This creates a natural link across the areas of the Internet of People, where people communicate and interact with each other or with "the Internet", and the Internet of Things (IoT) with sensors and actuators, and machine intelligence all connected to serve (hopefully) the needs of humans.

It starts with the core, the body, with body area sensors, continues with anything wearable, and beyond to anything owned, within physical or virtual reach.
The data ownership could be shared by a group, call it social, with a defined aim, for instance producing CBPP (Common Based Peer Production) as in the world famous Wikipedia. Note that the P2P value project of the EU addresses the topic of organisation and mechanisms at play in CBPP, with over 300 such social groups studied.
An other interesting case of data ownership is the Industrial Internet, where companies generate for their own operation data, which they use internally (mostly), in schemes such as a supervised distributes system, using typically a control room. Today a telecom network, a transport network (railways in particular, but also metro, road air and sea transport), an energy network are in this category. Some subsets of such operations data may be eligible for the company to release it for specific use.

Data generated by wearable devices is also a category of interest to the business and consumer communities, with multiple purposes being envisaged already (sports, well-being, health, new forms of communications) and many more to come.


This can be one of the few large Internet brands. This can also be any company in operations such as the ones above. This can also be an individual aggregating their own data in multiple ways, for multiple purpose: current and future (forensic data, the extension of the collection of post cards and pictures into the general data domain).
Governments are data collectors. Organisations: public or private, acting in pursuit of business or social goals are data collectors.
Even when the data is accepted as not being subject to a price tag, its use must conform to established rules and laws.

A data collector builds consistent and structured sets from individual potentially unstructured data vectors.
This entails quality control of the source, or to use other language the "veracity" of the information. 
The aim is to prepare the input of an efficient data processing.


A data user is typically an organisation or an individual performing analytics on data sets. For this purpose they need to either directly collect data sets, or buy rights to access such dats sets for their defined purpose and scope from data suppliers, which are the data collectors, or data brokers acting on behalf of the data collectors (retail role).
Data users needs data sets suitable for their need. This is the demand-side in economic terms, and the data collector or data broker is the supply-side.

Note that the use of data through analytics may lead to decisions, with in turn such decisions producing data sets in the command domain, for remote and distributed execution of such commands implementing the decision taken.
For instance real time systems with a feedback loop, also called automated systems, or optimal control, do not only "observe the world" through IoT sensors, but they act on the world through actuators, and supervised control, typically with supervision in a control room, as explained above.
SCADA systems (supervisory control and data acquisition) are an important case of operational data use.

Naturally, when the data collector gathers data, and forms data sets, initial data D0 is transformed into D1 and the set accessed by a user for a specific purpose and scope is D1* (optimised or limited for this use).
Hence a data path from extraction, collection, shaping, homogeneising and fitting to user purpose.

Users purchase rights to use the data for their own purpose and scope, and payment flows possibly through the data collector, with part of the payment remunerating the data owners.
The organisation of payment and retail is being studied, and a publication addressing this subject is being prepared.

Copyright R. Di Francesco, 2015 

No comments:

Post a Comment