Introduction to Areal Interpolation and MGGGs maup Ruth Buck - PowerPoint PPT Presentation

Introduction to Areal Interpolation and MGGG’s maup Ruth Buck

Framing the problem 10 10 10 10 ? ? 10 10 10 10 10 10 10 10 ? ? 10 10 10 10 We have data for the blue source units… …that we want on the orange target units.

Given data for a set of subareas, we want to find a function that best estimates the whole surface so that we may predict values for a different set of subareas (Lam 1983).

Example 1: Centroid approach 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10

Example 1: Centroid approach ? = 20 ? = 60 ? = 50 ? = 30

Example 2: Overlay with simple aggregation 10 10 10 10 ? = 20 ? = 60 10 10 10 10 10 10 10 10 ? = 50 ? = 30 10 10 10 10

Example 3: Overlay with areal weighting 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10

Example 3: Overlay with areal weighting .23 .77 1 10 1 10 10 10 .77 .23 .23 .09 .67 1 .99 10 10 10 10 .68 .25 .08 * .01 .16 .58 1 .55 .45 10 10 10 10 .84 .42 1 .55 .45 1 1 10 10 10 10

Example 3: Overlay with areal weighting 2.3 7.7 10 10 ? = 22.3 7.7 2.3 2.3 .9 ? = 54.9 6.7 9.9 10 6.8 2.5 .8 .1 1.6 5.8 10 5.5 4.5 8.4 ? = 40.3 4.2 ? = 42.5 10 5.5 4.5 10 10

Which interpolation method is best? It depends on… • The data you are interpolating • Availability of ancillary data • Computational power • Time CIESEN 2018: Gridded Population of the World (~1 km)

Which interpolation method is best? It depends on… • The data you are interpolating • Availability of ancillary data • Computational power • Time NLCD 2011

Which interpolation method is best? It depends on… • The data you are interpolating • Availability of ancillary data • Computational power • Time Piotr Jaworski

Which interpolation method is best? It depends on… • The data you are interpolating • Availability of ancillary data • Computational power • Time OpenClipart

When is this relevant in redistricting? • We have population and other demographic data at the census block level that we would like at the precinct level

When is this relevant in redistricting? • We have election results at the precinct level that we would like to disaggregate to census blocks

When is this relevant in redistricting? • Absentee results are only available at the county level and our analysis uses precincts as the unit of analysis

When is this relevant in redistricting? • CVAP data is available at the census block group level and we need it at the level of census blocks or precincts

maup Package assign intersections prorate Uses the common Essentially a spatial refinement to join; maps each Overlays the source interpolate data source unit to the and target units and from the sources to target unit that outputs the targets based on contains the common refinement user-specified majority of its area weights