GeoDa for path calculations?

[Julia Koschinsky]

on behalf of Geoff Wise (please cc him in your response).

Thx: Julia

Julia Koschinsky, Ph.D.

Executive Director and Senior Research Associate

Center for Spatial Data Science

University of Chicago

https://spatial.uchicago.edu/

---

From: spatial...@lists.uchicago.edu <spatial...@lists.uchicago.edu> on behalf of Geoff Wise <geof...@gmail.com>
Sent: Saturday, November 20, 2021 1:18 PM
To: spa...@uchicago.edu <spa...@uchicago.edu>
Subject: [spatial] GeoDA for path calculations?

 

Hello, my name is Geoff Wise and I am familiar with Python but no experience in spatial analysis.  I have just started to explore your GeoDa site to see if it will be useful for my anti-gerrymandering research.  My research will leverage a people-centered (vs. geography-centered) description of spatial distributions of voters.  Given the wide variation in population density, and its correlation to partisan support, I believe that a people-centric approach to spatial clustering analysis is a better way to understand and optimize the redistricting fairness problem.

    For this work, I need to compute the "people distance" between voters in one precinct to the voters in every other precinct in the state.  I can then run spatial correlation analysis (e.g. spatial lag model) on a people-distance basis vs. a physical distance basis.   To do this, armed with the voting-age population in each precinct, I believe I need to:

1) Determine the geographic center (centroids) of a state's voting precincts from e.g. Redistricting Data Hub shapefiles

2) Draw a straight line between each pair of centroids, then determine the length of each line segment within the precincts on the intervening path between these two endpoint centroids.  I can then compute the people distance by weighting each segment length along this path by that intervening precinct's population density.

I believe that #1 is straightforward using Geopandas once I can figure out how to install that on my Windows machine.

  For #2, I could imagine digitizing the line as a series of points, then somehow determining which precinct encompasses each of those points.  Or perhaps someone has already coded such an algorithm.

Any advice?  I would be happy to list anyone who can help me with this technical aspect as co-author on my planned paper.

Geoff Wise

geof...@gmail.com / 1-513-476-2480

[Nicolas Cadieux]

Hi,

Geoda will be useful for the analysis.  I suggest you use Anaconda Python for the geopandas installation.  It’s very strait forward.  Install geopandas in a new python 3.8  environment.  QGIS, is probably the best software for creating and managing the data sets.  Output files from geopandas dans QGIS are compatible with Geoda.  

Nicolas Cadieux

https://gitlab.com/njacadieux

Le 21 nov. 2021 à 11:36, Julia Koschinsky <jkosc...@uchicago.edu> a écrit :



--
You received this message because you are subscribed to the Google Groups "Openspace List" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openspace-lis...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openspace-list/SJ0PR11MB50064001286039F36AB9F21CCD9E9%40SJ0PR11MB5006.namprd11.prod.outlook.com.