Design Rationale

Process of Designing the Dashboard

We started this dashboard by identifying data sources that could provide comprehensive eviction records across New York City over an extended time period. We gathered data from the NYC Open Data Evictions dataset, which contains complete records of all residential evictions executed by city marshals since 2017, including execution dates, addresses, and geographic coordinates. This dataset was crucial because it provides the granular detail needed to analyze both temporal patterns and geographic distribution of evictions. The timespan from 2017 to present was particularly valuable because it captures both pre-pandemic baseline rates, the dramatic impact of COVID-19 eviction moratoriums, and the post-protection recovery period.

After we established our data source, we did a deep dive into the records and determined what data stories we wanted to tell. We recognized that evictions needed to be examined from multiple angles: the temporal pattern showing policy impacts, the geographic disparities in displacement pressure, and the distribution of eviction rates across neighborhoods. We developed three visualization types that work together to answer these questions: small multiples for temporal comparison across boroughs, normalized rates for fair geographic comparison, and a histogram for understanding the full distribution of neighborhood experiences. We wanted this dashboard to convey that evictions aren't random events but concentrated crises that affect specific neighborhoods disproportionately, and that temporary policy interventions can prevent displacement but don't solve underlying affordability problems.

Our dashboard incorporates three distinct visualization approaches to provide a comprehensive view of eviction patterns across NYC:

Small Multiples (Temporal Line Charts): The first and most prominent visualization we chose was a small multiples design showing evictions over time for each borough as separate line charts stacked vertically. We wanted to show both the shared temporal pattern across all boroughs (the COVID-19 dip and recovery) and the distinctive characteristics of each borough (absolute eviction levels and rate of recovery). Small multiples are ideal for this because they allow direct comparison through aligned axes while keeping each borough's data clearly separated and readable. Each chart uses time as the continuous variable on the x-axis (2017-2026) and eviction count as the discrete variable on the y-axis. The line chart format emphasizes temporal continuity and makes trends immediately visible through slopes and the principle of continuity, where the connected line creates a clear visual path showing the trajectory of change. We maintained our consistent borough color scheme (Bronx blue, Brooklyn orange, Manhattan red, Queens teal, Staten Island green) to create visual continuity with the other dashboards and leverage the principle of similarity so viewers naturally associate colors with specific boroughs. The synchronized x-axes are crucial because they ensure that the dramatic valley in 2020-2021 appears at the same horizontal position in all five charts, making the shared policy impact unmistakable. The varying y-axis scales accommodate the different magnitudes across boroughs while maintaining proportional representation within each chart, adhering to data-ink ratio principles by eliminating unnecessary gridlines and decorative elements.
Choropleth Map (Evictions Per 1,000 Households): The second visualization is a choropleth map showing evictions per 1,000 households by neighborhood, which uses a sequential pink color scale to indicate displacement pressure across predefined zip code areas. This map displays ratio data (evictions normalized by population with a true zero point) through color intensity encoding, where darker pink shading represents higher eviction rates. We chose pink as our color scheme to distinguish this dashboard visually from the red used in rent burden while still maintaining a warm color that conveys urgency and concern. The map format allows viewers to identify geographic hotspots of displacement, with the darkest concentrations revealing where eviction pressure is most severe. The sequential color scale creates natural grouping through the principle of similarity, where neighborhoods with comparable eviction rates appear visually alike, while the principle of proximity helps viewers perceive adjacent neighborhoods with similar shading as related geographic clusters facing similar displacement challenges. The normalization by population is critical because it shows eviction pressure relative to neighborhood size rather than just absolute counts, making it possible to identify areas where residents face the highest risk regardless of total population. The choropleth format adheres to data-ink ratio principles by using the geographic boundaries themselves as the primary structural element, eliminating the need for gridlines or additional decorative features. The figure-ground principle creates strong contrast between the darkest pink neighborhoods (the figure) and lighter areas (the background), immediately drawing attention through the focal point principle to the areas facing the most severe eviction pressure and making geographic disparities unmistakable.
Histogram (Distribution of Neighborhood Rates): We recognized that borough-level patterns and even the map hide the full range of neighborhood experiences, so we wanted to show the complete distribution of eviction rates across all neighborhoods. Therefore, we chose a histogram that groups neighborhoods into bins based on their evictions per 1,000 households rate. This visualization transforms continuous data (eviction rates) into a frequency distribution using bins as intervals on the x-axis and frequency counts as discrete values on the y-axis. The histogram reveals the shape of the underlying distribution, showing that most neighborhoods cluster in the lower ranges but there's a long tail of neighborhoods with extremely high rates. We used yellow for the bars to create strong contrast against the dark background through the figure-ground principle and to differentiate this chart visually from the borough-colored visualizations. Yellow also carries a cautionary connotation that fits the concerning nature of the data. The bin widths create groupings through the principle of common region that make patterns visible, such as the concentration of neighborhoods in the 40-70 range and the presence of outliers above 250. The histogram format is ideal for showing distribution shape because the area of each bar (height multiplied by width of the bin) represents the frequency of occurrences within that range, making it easy to see where most neighborhoods fall and where the extreme outliers exist. This visualization maximizes the data-ink ratio by presenting only the essential frequency bars without chartjunk or unnecessary embellishment.

The dashboard layout uses small multiples as the dominant left-side element because temporal change is central to understanding evictions in the context of COVID-19 policy impacts. Placing all five borough charts vertically with aligned axes creates an intuitive scrolling reading pattern that emphasizes both individual borough stories and cross-borough comparisons. The right side presents two complementary views: the normalized comparison that corrects for population differences and the distribution that reveals within-borough variation. This layout moves from temporal (left) to cross-sectional (right) analysis, and from aggregate borough trends to granular neighborhood distributions.