The following visualizations depict several facets of data concerning New York City’s “Leading Causes of Death” from 2007 to 2011. The dataset used to create the visualizations (using Tableau Public) in this post can be downloaded from NYC OpenData; the data was collected by the Department of Health and Mental Hygiene (DOHMH).

*Click links [Viz 1, etc.] for interactive visualizations

Viz 1: What are the top two causes? 

Notice I’ve isolated heart disease and tumors in this viz–this is because I’ve predetermined them to be the top causes of death (by quite large margins) for all ethnicities and both sexes from 2007 to 2011. Across almost all ethnicities except for Asian & Pacific Islanders, more women died from heart disease than men while more women died from tumors than men with the exceptions of Hispanic and Asian & Pacific Islander men. The amount of deaths relating to tumors remains consistent throughout the years, while heart disease steadily decreases (though, from 2009 to 2010, there is a a bigger dip than other year ranges [19,803 -> 17, 687]). The red gradient used for the horizontal bar graph is a cognitive tool—it illustrates that a greater amount of deaths equal a darker shade. A legend denotes the sexes using the gendered colors of pink and blue.

Viz 2: Leading causes of death by ethnicity and sex

Here are the rest of the leading causes of death by ethnicity and sex sans top two causes. I decided to remove the top two causes of death from the default viz because it renders a smaller range of values to work with for better examination of data. By using pink and blue circles without a fill as the plot points, a user can see where women and men have had similar amounts of deaths (illustrated by overlapping circles; see image below). It’s instantly apparent that the highest amount of deaths (~3,300) in this viz belong to Caucasian females who suffered from influenza and pneumonia.

To highlight a particular cause of death, it can simply be clicked on in the y-axis; this will isolate the data of interest, fading out the rest of the data. Alternatively, because there is such a wide range of values and numerous causes of death, the Quick Filter (which includes a checklist and a search option) can be used in order to more closely examine particular causes (or compare causes) and their corresponding amounts of deaths—pretend you’re taking a magnifying glass to the data you are most interested in. That’s what Quick Filter does.

Viz 3:  Total amount of deaths over time

Let’s look at the big picture: total amount of deaths over time. There is a staggering amount of deaths in the Caucasian population in comparison to Africans, Hispanics, and Asian & Pacific Islanders.  The amount of deaths for all ethnicities seem quite consistent over the five years i.e., there is no outstanding increase or decrease of deaths for any year. More women died than men over the five year period. In 2010, women and men had the smallest difference in amount of deaths (20,086 vs. 19,761)—though women still, sadly, triumphed here.

Amount of Deaths over Time (by Ethnicity) and Amount of Deaths over Time (by Sex) show scales of 0 to 25,000 and 15,000 to 25,000, respectively. Amount of Deaths over Time (by Sex)’s lowest plot point is ~19,700, so I truncated the y-axis range from 0-25,000 to 15,000-25,000 to allow for a closer view of the data. (See images below).

 

Viz 4: What are the leading causes by ethnicity? / Viz 5: What are the leading causes by sex?

These two highlight tables, while not very exciting-looking, are quite useful for analysis of data. Firstly, the tables immediately shed light on gaps in the data. For example, looking at Causes of Death (by Sex), it can easily be discerned that there were no recorded deaths from viral hepatitis among women; looking at Causes of Death (by Ethnicity) there is no recorded data for congenital malformations, deformations in Asian & Pacific Islander, Caucasian, and Hispanic populations. The visualizations can also highlight the lowest amounts which can sometimes be lost in a dataset with a large range of numbers: e.g., Causes of Death (by Ethnicity) makes it easy to see that aortic aneurysm and dissection had the lowest amount of deaths amongst the four ethnicities.

In the same vein as Viz 1, the red gradient used for these two highlight tables allows a user to easily see that a darker shade equal more deaths. These two visualizations can be used as an alternative to Viz 1 which displays all data (disregarding years); however, these visualizations provide a clearer overview of the total amounts of deaths for each corresponding cause.

–Sarah Hatoum