1

I'm working on a study on gentrification using ArcGIS, and I want to see which areas in my city have seen major increases or decreases in property value over the last few years (and map that to changes in demographics, but that comes later). Each year's property appraisal data has shapefiles for the boundaries of each lot and a table with the value of each lot (which I've managed to join, though a few lots in the shapefile have no entries in the table for whatever reason; I'm talking to the appraisal people about that). However, the lot boundaries aren't always consistent (due to properties being subdivided or combined), and I'm not sure if IDs for a given property are always the same between years.

I want to create a layer that shows changes in value from 2009 to 2010, 2010 to 2011, etc, as well as total change from 2009 to 2014. I'm not sure how to deal with changes in property boundaries; maybe if a property is divided during that period, also apply the division to previous years and divide the value either by the area (which assumes that property value is uniform) or proportionally based on the newly divided lots' values (which might result in the effects of later development being applied ahistorically). How would I do all this?

  • If you're not expecting to get down to a very fine level and only want a generalization of the whole city, province or country then perhaps rasterize with a cell size of 100m or thereabouts, that way you only need join within a year and you can use raster calculator.. A tool like Euclidean Allocation will help fill in the blanks inbetween where lots haven't been valued (or are missed), it is also very quick and easy to see how the value changes between years and decades with raster functions. – Michael Stimson Mar 30 '15 at 5:32
  • I'll try that. I'm looking to generalize it to the census tract level. – sbl Apr 1 '15 at 16:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.