Skip to main navigationSkip to main content
The University of Southampton
Web Science Institute

Modelling Data Workflows to Understand Delivery of Data Resilience


The proliferation of IoT devices has generated a deluge of live and historical sensor data which could be useful to diverse stakeholders in diverse settings. However, this IoT data can often be difficult to collect and use because it is fragmented, inconsistent, and lacking information relating to quality. The challenge is therefore to establish greater data resilience, including data management, data quality, service continuity, redundancy, assurance, and trust. This challenge is recognised in Innovate UK’s funding programme “Addressing cyber-security challenges in the Internet of Things” which identifies data resilience as a strategic focus.

Proposed solutions for data resilience have focused on service or discipline specific models where standards can be controlled[1] or on primary application of data to specific settings. However, there remains a disconnect between data management and application across diverse settings. To unlock the potential of IoT data for diverse stakeholders, a cross-disciplinary approach to data resilience is needed. Such an approach should address the disconnect whilst minimising impact on the management of primary data application.

One approach to improving data resilience involves using a metadata-based architectural model for dynamically resilient systems[2]. A second approach could use workflow analysis techniques, such as 4D modelling[3], to model the spatiotemporal qualities of data pipelines. In collaboration with the GeoData Institute, we will use these approaches to model the workflows and roles of stakeholders in a service where a diverse range of geospatial data are utilised from a range of data providers to establish generalisable approaches to delivering data resilience. Our approach will be informed by outputs from analyses of stakeholder concerns which emerged from the UoS-CISCO Data Sharing Policy and Standards Workshop series, the PEDASI Project and the Big Marine Data Project[4]  .

Analysis will comprise:

  1. Reviewing data surfaced via the Big Marine Data project to understand common features providing an indication of data quality, e.g. license, api/other, an applied schema standard;
  2. Modelling the required components in the data workflow to deliver data resilience, in particular via assessment of data quality;
  3. Undertake a workflow analysis based on a use case for a data driven service using IoT data, giving particular consideration to the impact of delivering data resilience for stakeholders within the workflow e.g. implementation of  a quality assessment tool.

The Big Marine Data Project (Blue Marine Foundation) delivered by the Geodata Institute established the use of a web observatory in demonstrating data aggregation potential around a discipline specific subject; observations include RNLI rescue statistics, metocean real-time sensor measurements[5], and sea state models[6]. Through this stimulus project’s partnerships we will use data provided by RNLI and Lloyds Register to explore the application of techniques which can deliver resilience within the workflow to provide confidence in systems utilising such data aggregations.

Utilising modelling techniques, a workflow diagram will be developed to demonstrate the process, dependencies and technology/policy gaps in the data pipeline, identifying and prioritising interactions that will have the greatest potential to improve data pipeline resilience. Following this, we will establish a Quality Assessment Criteria for data based on analysis of quality, legal and economic decisions made by stakeholders throughout the data pipeline. Finally, alongside input from domain experts, we will evaluate findings from these tasks to inform the development of an outline specification for a mobile/web app for viewing and analysing marine safety data. These outputs will provide a major component of proposed follow on research and development funding.

[1] Internet of Things for Smart Cities: Interoperability and Open Data

[2] Seugendo et al,

[3] Developing high quality data models


[5] Channel Coastal Observatory,

[6] Plymouth Marine Laboratory FVCOM modelling,



Principal Investigator: Professor Dame Wendy Hall

Co Investigators: Jason Sadler

Privacy Settings