Breaking Up Multimember Districts

---

In Merrimack, NH, we addressed the challenges of multimember districts, where multiple representatives are elected at-large—a system that can dilute the principle of one person, one vote and often favors one political party. 

Using block-level census data, we developed a Flow-Based Partitioning Algorithm that redistricts the town into eight balanced wards (left). The algorithm works iteratively to create contiguous and compact areas with equal populations. To further contextualize the maps, we overlaid additional data such as current representative addresses and potential polling locations, making it easier for the public to understand and support the changes. 

This data-driven approach provided a scientifically sound solution to redistricting and delivered a unique, timely, and efficient service that no other entity in the state could match.

We analyzed the use of NH RSA 644:6, the statute that criminalizes loitering and prowling in New Hampshire. While the law is intended to maintain public order, community members have raised concerns that it disproportionately targets unhoused individuals. 

To explore these claims, we collected arrest records and census data from towns and cities across the state and developed a Poisson log-link regression model (the blue line). This model predicts the total number of arrests we should expect in 2022 (y-axis) based on the population size of each community (x-axis). By comparing these predictions to the actual arrest figures (the dots), we identified the top three towns or cities with the highest excess arrests (the labeled, red dots).

This analysis grounded the project in rigorous data science and the scientific method, moving beyond anecdotal evidence. The insights gained not only enhanced the efficiency of the project but also helped reduce costs through targeted resource allocation for future initiatives.

We processed and visualized 2024 election results for New Hampshire State House of Representative elections in cities such as Manchester, NH. 

The visualization shades each of the twelve (12) city wards based on the closest margin between Democratic, Republican, and Independent candidates. The darkest color indicates the most competitive districts; as few as nine (9) votes made the difference between winning and losing. Dots on the map represent the number of candidates elected per ward (in this case, two), and their color represents the winning party in 2024. 

Election results in these instances can be complex —especially in multimember districts with multiple candidates per party. This project transforms that complex dataset into clear, easily digestible visuals that answer common questions about New Hampshire’s electoral landscape. These dynamic visualizations empower political campaigns to make data-driven decisions, optimize resource allocation, and fine-tune campaign strategies.