Network analysis for big data

Crispin gave an invited talk at Association for Geographic Information’s “GeoBig5: Big Data and You” conference on the links between spatial network analysis and big data.

To summarize,

  • Spatial network analysis leverages existing large datasets by extracting novel information from them.  The mere shape of a road layout can tell us volumes about land use, travel demand and preferred routes.
  • As every, nobody can quite agree on what Big Data is!  (If computing these days is Big Data, does that make maths Big Sum? Thank you Amanda Turner for this quip).  Almost anything in the GIS field fulfils the “variety” requirement of big data though, and sDNA is no exception as it integrates with GIS to unify diverse data sets.  Examples include our work on community cohesion and use of sDNA in the UK Biobank big data project.
  • sDNA is geared towards high computing requirements; e.g. a 300,000 link network contains 90 billion potential interactions, each of which require tracing a route from origin to destination.
  • As much of human interaction is mediated by spatial networks, ignore them at your peril!

AGI members can download the presentation from their website.