28 November 2016: Malcolm Atkinson EVENT POSTPONED (new date tbc)

Our abundance of data from multiple sources and the commercial, research and societal challenges we try to address require collaborative behaviour: pooling information, methods and resources. I used to think of this as a technical challenge: either platforms could not support the integrated and holistic view or such platforms could only be afforded by dramatic causes. But today, the technology and platforms have reached the capacity and affordability to expose a social issue – “the owners do not want to share”.

Data-Intensive Federations are those that fall between the “everything is open” and “everything is mine” extremes. The participants retain rights, have commercial interests, are governed by stakeholders with expectations, and operate in different legal frameworks. Such DIF are created and sustained to meet a cluster of targets and may impose their own ethos and governance. Meeting these complexities in a reliable way requires computer-assisted automation; its form and nature are a proposed area of research. It has to be trusted by all participants in a heterogeneous distributed digital world. A crucial component will be support for “Data Diplomats” who negotiate compromises between stakeholders with differing and strong “group identities” to make a DIF feasible and effective. The talk will use case studies gathered over 40 years to expose what is involved and what is not involved in overcoming the socio-political impediments so that data science can work for multi-stakeholder systems.