Gerrymandering is a growing national concern. Prior to the 1960s, Congressional districts often respected county and municipal boundaries, but wildly differing populations among districts led to the establishment of the one person, one vote standard. In an effort to conform to the requirements of one person, one vote, Congressional districts became increasingly less compact, and increasingly more gerrymandered. With cheaper computing power and the increasing adoption of GIST, the redistrictings implemented by several states following the 2000 and 2010 censuses fare poorly on measures of compactness and have arguably been among the most egregious gerrymanders the United States has ever seen. In different states, gerrymanders benefit both Republicans (Pennsylvania, Wisconsin, North Carolina) and Democrats (Maryland, Illinois). In the face of increasingly obvious gerrmanders, several court cases have successfully challenged Congressional districts in lower courts, and some of them have been taken up by the Supreme Court for review.
In this talk, I discuss the use of open data and open source software (primarily R and QGIS) to analyze redistricting and model election outcomes of hypothetical districts. Additionally, I discuss my work with Concerned Citizens for Democracy, a nonpartisan Pennsylvania-based advocacy group. Pennsylvania, like several other state constitutions, requires that counties and municipalities be split only if "absolutely necessary". We demonstrate that it is possible to achieve equal district populations with many fewer county and municipal boundary splits than in the current, recently invalidated PA district plan. I will also discuss redistricting guidelines that, if adopted by the courts, will severely limit the ability of policital parties to create gerrymandered districts.