Aggregation and interpolation

User Requirements RAGLD Component(s) RAGLD Considerations
Aggregation and interpolation
  • Geography to geography (one set of abstract areas into another)
  • Enriching one kind of dataset to make it interoperable with another.
  • Point data to spatial regions.
  • Dataset expressed in region typeX but want to view as typeY.
  • An enactment of statistical transformation operations.
Issues where two region types spatially overlap, but neither a subset of the other.
E.g. “I have population for wards but want to find population by school catchment area.”
Convert between different levels of geography
E.g. “Give me all the deaths in Hampshire if I know the deaths in all of the settlement regions”.
Dataset transformation services
  • This may have difficult aspects. Creating higher level aggregations from smaller regions is easiest, but typeX and typeY may not be direct subsets of one another (overlap, etc).
  • Producing smaller areas from larger regions requires interpolation – based on area? density? population? something else?
  • These are generally expensive off-line computations, not done on-the-fly.
  • Likely to be a mix of on-the-fly and offline/batch services.
  • This may have difficult implications, aggregation may be easy, interpolation hard?
  • Can we make this a generic service or not?
  • Could use ONS method of population weighted centroids?