Three New York Times visualizations

Mapping Foreclosures in the New York Region

Mapping Foreclosures in the New York Region

Given that the New York Times Graphics Department is a winner in this year’s National Design Awards, it seemed opportune to look back at some of its recent work. Over the past few years, the Times has published many excellent interactive visualizations as counterparts to the equally brilliant static information graphics found in the paper, including the previously mentioned 31 Days in Iraq by Alicia Cheng. Each interactive is predicated upon a hypothesis and the evidence that supports it. Here, visualization is treated as a medium for journalistic inquiry by creating an editorial framework for the data on display.

Through its relationship to published news articles in the Times, every piece has a specific agenda and questions it seeks to find answers to, which builds immediate engagement with the viewer. A New York Times visualization generally summarizes an issue that has been reported on over a stretch of time, offering analysis and insight while it is still noteworthy. Most visualizations are attributed a specific date and represent a data-snapshot of a particular moment in time. The multivariate data and open-ended modes of interaction allow viewers to ask further questions of the issue and find answers through exploration.

Three of these visualizations have struck me as particularly successful. One of the most recent maps foreclosures in the New York region.

Foreclosures in the New York region (zoomed-in to city level)

Foreclosures in the New York region (zoomed-in to city level)

The premise of this piece is that foreclosures were most frequent in those areas of the New York region with large minority populations, clearly shown by plotting each foreclosure, aggregated by individual buildings, on a map of the area, in addition to showing the percentage of foreclosures per residential units per census tract. The map makes use of three levels of detail in accordance to zoom-level. At the highest zoom-level, only the census tracts are visible. Zooming in, individual foreclosures by building form clusters around neighborhoods and major roads.

Foreclosures in the New York region (zoomed-in to street level)

Foreclosures in the New York region (zoomed-in to street level)

At the lowest zoom-level, the granularity increases to show individual streets, blocks and intersections. According to its premise, the visualization highlights the neighborhoods most stricken by the housing crisis, and it turns out that they are indeed areas with high minority populations, including Bushwick in Brooklyn, Jamaica, Queens, and Newark, NJ. In addition to retrieving the exact percentages on hover, one can also move between 2005 and 2009 to see the change in foreclosures over time.

The visualization method takes after Dr. John Snow’s map of the London cholera epidemic and while (like the latter) it is uncertain whether the visualization served to formulate a hypothesis or to validate it, it is nonetheless a very effective use of a geographic scatter plot. Improvements might include more granular hover states to the building level, and aggregate numbers for not only for census tracts and cities, but also neighborhoods and boroughs.

A Year of Heavy Losses is a treemap visualization showing the change in market capitalization of 29 selected financial companies on Wall Street over the course of a year (October 2007 to September 2008).

A Year of Heavy Losses (October 2007)

A Year of Heavy Losses (October 2007)

A Year of Heavy Losses (September 2008)

A Year of Heavy Losses (September 2008)

What makes this visualization successful is its simplicity. The treemap format effectively shows the relative sizes of market cap, and the before/after comparison makes the overall sector loss abundantly clear—with rather brilliant subtlety, the boundary of the October 2007 visualization remains visible in the September 2008 comparison. Through color one is also able to compare five types of financial company: Investment banks, national banks, financial services, regional banks, and asset managers/investors. Finally, the toggle used to switch between years also depicts the size of financial sector in relation to the total stock market. Overall, this is an effective and elegantly compact information graphic.

Casualties of War, finally, offers an on-going analysis of US military fatalities from the Iraq war, by location of death, age, race, military branch, and type of duty.

Casualties of War (initial invasion)

Casualties of War (initial invasion)

Casualties of War (since troop buildup began)

Casualties of War (since troop buildup began)

Casualties of War (home states)

Casualties of War (home states)

Mike Migurski of Stamen Design reviewed this piece in early 2007. I won’t add much more apart from summarizing the aspects that make it an effective visualization, from the flexibility of the timeline, which not only allows the variable selection of time-ranges, but also charts the number of casualties, to the side-by-side display of multiple data records and the inclusion of preset key milestones. Its data rich presentation allows open-ended exploration with a seemingly infinite number of possible time-ranges.

The characteristically understated (one might say transparent) presentation—showing difference through the smallest possible stylistic variation—respects the gravitas of the subject matter, and the only hint of political affiliation (apart from the choice of subject matter in itself) is the terminology used within the preset milestones. Rounding out the core analysis are stories of the fallen and an expressive visualization of each service member who died in Iraq, represented in a grid in order of death—a virtual memorial.

Following Tufte’s principle of data density, these three interactive pieces, among many others published by the Times, are exercises in the effectiveness of information visualization as a technology, for both sense-making and storytelling.

Comments are closed.

related posts

elsewhere