Presented By O'Reilly and Cloudera
Make Data Work
Feb 17–20, 2015 • San Jose, CA

How Minority Becomes Majority - A Study of Gerrymandering

Tatsiana Maskalevich (Stitch Fix)
1:30pm–2:10pm Thursday, 02/19/2015
Law, Ethics & Open Data
Location: LL21 B
Average rating: ****.
(4.00, 3 ratings)

The term “gerrymander” was first introduced by the Boston Gazette in 1812 as a response to the redistricting in Massachusetts under then-governor Elbridge Gerry. One of the newly drawn districts from the infamous redistricting bill signed by Gerry resembled the shape of a salamander. By combining the governor’s name with salamander, one arrived at the term gerrymander. Recently, the great American political theorist Jon Stewart, and his team of analysts at Comedy Central’s “Daily Show,” have offered gerrymandering as an explanation for political gridlock and government shut downs. How has the method for drawing a congressional district impacted our government in the US?

The Daily Show analysis is compelling! But is it true that dominant parties are drawing themselves safe districts? Unfortunately, our analysis of the evidence forces us to conclude that the Daily Show is wrong! Republicans don’t draw overwhelmingly safe districts for themselves, the do so to democrats, and vice versa! The party drawing the map can actually engineer a congressional majority for themselves by drawing an overwhelmingly safe district for their opponents!

To go from truthy to true, we wanted to undertake a rigorous analysis of gerrymandering. In this talk we will discuss our approach to quantifying gerrymandering of a congressional district. We started with the ratio of the shape area to its perimeter, called “compactness.” If a gerrymandered district is “squiggly,” can we use compactness to understand the extent to which the district is drawn to have specific demographic characteristics?

We then turned to our exploration of partisanship. We deployed Partisan Voting Index data to identify swing districts and districts with one-party leaning. Voting, however, is not a great indicator of the partisanship of a certain representative. Most votes occur substantially along party lines, so the voting behavior obscures the ideological leanings of the representatives. We will discuss our exploration of understanding the funding and positions of representatives, rather than just their voting behavior, to understand how polarized the delegations have become.

It turns out that our friends in the late night comedy business may have missed the most scandalous result of gerrymandering. The most surprising impact gerrymandering can have is to create a congressional delegation that is at odds with the electorate of a state. For instance, in 2012 in North Carolina, the statewide vote was 51% for Democrats and 49% for Republican congressional candidates. However, the resulting congressional delegation was made up of only 4 Democrats, and 9 Republicans! We will expand on the analysis from the opinion page of the New York Times and discuss in depth our analytical observations on the practice.

Photo of Tatsiana Maskalevich

Tatsiana Maskalevich

Stitch Fix

Tatsiana Maskalevich is a data scientist at Silicon Valley Data Science.
Blending both industrial and academic research, Tatsiana is expert at solving hard business problems. She brings a background in both mathematics and statistics, and has deep experience researching and implementing models for predicting user behavior.