Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Tom Heath: Data architecture: from smoke alarms to towns and cities

Tom Heath: Data architecture: from smoke alarms to towns and cities

This presentation supports Tom's talk about data architecture, which he presented at Swirrl's 'Data-Driven Decisions' event in Manchester, May 2106.

Swirrl

May 26, 2016
Tweet

More Decks by Swirrl

Other Decks in Technology

Transcript

  1. Data architecture: from smoke alarms to towns and cities Dr

    Tom Heath Data and Systems Architect Arup Digital [email protected] @tommyh Manchester, 26 May 2016
  2. 2 Context and objectives •  Previously Head of Research at

    the Open Data Institute •  Now Data and Systems Architect at Arup •  Arup and the digital built environment •  Exploring data silos across the built environment, at four levels of granularity •  This talk as a test of some evolving ideas - feedback please, new ideas, new connections - (and a beer for whoever spots the deliberate error in my slides)
  3. 4 Photo credit: Bob Jenkins, flickr:48380660@N04 •  Critical domestic infrastructure,

    often poorly maintained •  Source of high value data •  Low volume, low velocity and very analogue! •  Want to live the home automation dream? Install our app! •  No! •  What is the appropriate architecture for domestic data? Smoke alarms
  4. 5 •  How many maintenance contractors does it take to

    change the light bulbs? •  00s, 000s of devices, highly complex •  Data and systems typically silo’d and proprietary •  BIM: Building Information Management •  Buildings as databases, but not just bigger silos •  How does data flow around a building? How is it interconnected? Is every device an endpoint? •  What is the appropriate data architecture? Buildings Photo credit: Brad Greenlee, flickr:bgreenlee
  5. 6 •  Scale up to a neighbourhood or district • 

    Where is the data? •  Relevance to me is not evenly distributed across a town/city •  Is data appropriately contextualised data consumers? •  Disconnect between hyperlocal vs administrative viewpoints - bounding criteria: does your lens as a resident match that of your local authority? Neighbourhoods Photo credit: Axel Naud, flickr:axelnaud
  6. 7 •  Data management practice is defined by administrative structures

    •  Should the same be true of data publishing? •  Are (most) (open) data portals just creating new (duplicate) silos? •  How does data flow around a town/city? •  How is it interconnected? How should it accessed? •  What is the appropriate data architecture? •  Are all actors properly represented? Towns and cities Photo credit: Brian Marble, flickr:lostmahbles
  7. 8 •  Can the same data architecture scale up from

    buildings to cities? - Many common features and requirements •  Multiple writers, multiple readers - design for appropriation and unintended reuse •  Multiple data owners - favour openness to reduce friction •  Many-to-many connections between data points - design for data connectivity •  (Grab me later for the technical bit) From buildings to cities
  8. 9 Urban data platforms: more akin to content management systems

    or search engines? Discuss. Take home thought