A Visualization of Nitrogen Dioxide and Asthma Risk by District in 2018
About This Poster
This poster explores where air quality in New York City still fails to meet basic safety standards—and how that overlaps with respiratory health risks. While nitrogen dioxide (NO₂) pollution has declined across the city over the past decade, the improvements haven’t been equal everywhere.
Using 2018 data, this visualization compares NO₂ levels and adult asthma episode rates at the community district level. It highlights the five most and least polluted districts and tracks how their air quality has changed over time. The goal is to show how environmental and health burdens can persist, even in a city that’s getting cleaner overall.
How I Got Here
1. Starting Confused—But Curious
I felt completely lost when downloading the NYC Air Quality Surveillance Data from data.gov. I wasn’t sure what most of the columns meant, especially the one labeled “Name,” which included several unclear terms. After some digging and engaging in helpful discussions, I discovered that the entries in the ‘Name’ column referred to different air pollutants. This was my first breakthrough.
Eventually, I identified four main types of pollutants: Ozone, Nitrogen Dioxide (NO₂), Fine Particles, and Boiler Emissions. This clarity helped me understand the structure of the data and provided a foundation to work from. I started by creating simple bar charts comparing pollutant levels by borough. However, since each pollutant used a different unit of measurement, I had to separate them.
Next, I developed trend charts focusing on NO₂ and ozone, as they shared a unit of measurement, allowing me to visualize their changes from 2010 to 2022. This became my initial dashboard.
After sharing my initial dashboard, I received feedback from my classmates and friends who regularly work with data as engineers. Here’s some of the feedback I received: “The Y-axis unit labels were unclear,” “Manhattan seems to be the worst – can that be highlighted visually?”, “Timeframes weren’t consistent or clearly labeled,” “What do the data values represent?” and “Could you summarize pollution overall by borough?”
Based on this feedback, I revised my dashboard. I learned that the time period is crucial for readers and that data design must depend on the contents of the dataset; I couldn’t force everything to match if the data wasn’t available. Additionally, I realized the importance of clear labels and consistent timeframes. Receiving feedback helped me see what wasn’t clear to others.
It was through this charting process that I started to crave spatial context. I wanted to understand where the air quality was the worst, so I decided to create a map. This exploration into spatial visualizations made my work much more engaging.
2. The visualization that makes more sense
While the chart and graph visualizations were helpful for gaining an overview of the dataset, I realized that there was a need for spatial visualization. Before diving into map visualization, I took a moment to reflect on the topic, what I wanted to communicate to my audience, and the story I aimed to convey. I decided to adopt a more audience-friendly approach.
First, I organized and clarified my topic with a more defined direction. Next, I included concise explanations for each pollutant, detailing their causes and the health issues they can create. Finally, I chose to focus my entire project on one pollutant: nitrogen dioxide. This choice was based on its significant effects on respiratory health and its suitability for storytelling.
This is my first map visualization. I joined air quality data with district boundaries from the NYC planning website. Since some names did not match, I manually added the missing borough and community district numbers to complete the join. This map displays the average nitrogen dioxide levels across New York City’s community districts in 2023.
The colors indicate the air quality in each area:
- 🟦 Blue represents cleaner air, with lower NO₂ levels.
- 🟧 Orange indicates more polluted air, with higher NO₂ levels.
The gray or blank areas on the map signify locations where there is insufficient data available for 2023. I chose to leave these areas blank rather than guessing the values, making the map more honest, even if somewhat incomplete. While it might look better with filled-in estimates, doing so would compromise accuracy. Despite some missing areas, the map effectively illustrates how NO₂ levels vary across the city.
We can observe the overall trend:
- Nitrogen dioxide levels are significantly higher in central Manhattan and lower in the outer boroughs, such as Queens, Brooklyn, and Staten Island.
In summary, this map helps us understand where the air is cleaner and which areas may need more attention in efforts to improve air quality.
For the second map, I aimed to address the feedback I received in the previous assignment, which suggested that segmenting the data could help tell a clearer story. Instead of using a gradient scale, I divided each district into two categories: Safe or Not Safe, based on whether the NO₂ level is below or above 20 ppb, according to the NYC Community Air Survey guidelines. This approach has proven beneficial and effectively draws attention to the issue.
3. Trying Network visualization
Initially, I aimed to create a network visualization for my NYC air pollution project, focusing on how major pollution sources affect neighborhoods. However, I found that the necessary source-to-district data wasn’t available for analysis. Instead, I decided to enhance my Gephi skills using the dolphins.xml dataset from CASOS at Carnegie Mellon, which captures social relationships among dolphins. Although it wasn’t directly related to my project, it served as a great learning tool, and the final visualization even resembled a dolphin’s silhouette, making the experience feel poetic.
As this practice is not highly related to the NYC Air project, I will keep it concise.
4. Getting Into Final Work — With Some Upgrades
As I continued working on the project, I realized that air pollution levels alone didn’t tell a complete or compelling story. Just showing where NO₂ was high wasn’t enough—I wanted to know: were these pollution levels actually affecting people’s health? That’s when I decided to add a public health layer to the map by incorporating asthma data.
I found asthma episode rates through NYC EpiQuery, which provides community-level health data. The most recent asthma data available was from 2018—conveniently, this matched with a cleaner, more complete version of the NO₂ dataset from the same year. So I pivoted fully to 2018 as the baseline for my final work.
Overlaying the asthma data revealed that several of the most NO₂-polluted districts also had higher asthma episode rates. While not perfectly aligned, the correlation was strong enough to support a more layered narrative. The addition of health outcomes brought greater depth to the overall story.
To explore the timeline, I selected the five most and least polluted districts and visualized how their NO₂ levels changed between 2009 and 2018. I paired each with a line chart, showing whether air quality had improved—and how quickly. The contrast was striking. Some districts crossed below the 20ppb safety line over time, while others stayed above it for the entire decade.
I also refined the structure of the poster:
- The first section asks the key question—Where is it still unsafe to breathe in NYC?
- The second shows how NO₂ and asthma burdens overlap.
- The third breaks down the “Top 5” and “Bottom 5” districts.
- The final visuals emphasize unequal progress through trends.
Throughout the design process, I continued getting helpful feedback. One of the most useful suggestions was to make chart titles more narrative—not just “what” the chart is, but “what it’s telling us.” I revised the titles to reflect key insights and added short captions beneath each visual to reinforce the story. This helped tie everything together into one cohesive arc.
5. Looking Back
When I first downloaded the dataset, I had no idea what I was doing. I just knew I wanted to learn how to make something with data—and maybe tell a story with it. At the beginning, I couldn’t even tell what the numbers meant. But step by step, I figured it out.
What really changed everything was realizing that pollution levels alone didn’t mean much unless we asked: “Who is this actually affecting?” That’s when I brought in asthma data and the project started to take shape. It became less about just numbers, and more about people.
Not everything worked the way I hoped. Some charts didn’t make sense. Some maps looked messy. I even tried a network visualization that didn’t fit the project at all. But I kept going, and with each mistake or revision, I learned something new—about tools, design, and what makes a message clear.
In the end, I’m proud of how far this idea came. It’s still not perfect, but it feels honest. And if someone else looks at the final poster and thinks, “Oh… I didn’t know the air was still that bad in some parts of the city,”—then I think it did what it was supposed to do.