While navigating the challenges of visual complexity in maps

Air quality has become a prominent topic in recent media discussions. For me, this term gained significant attention during the summer of 2023 when the sky over New York City turned an ominous shade of orange. I vividly recall the heavy smell of smoke lingering on clothing after even brief walks outside. Everyday activities suddenly posed health risks, prompting a sobering realization of the challenges faced by many individuals, particularly during wildfire season. The interconnection between wildfires, air quality, pollution, and emissions became starkly evident. Therefore, when I stumbled upon a dataset providing comprehensive insights into total passenger car sales across 141 countries from 2005 to 2022, I was intrigued. My immediate instinct was to seek out a dataset that tracked average global air quality, pollution levels, or car emissions. Fortunately, I discovered a dataset containing air quality measurements from 8,469 locations across 65 countries.

I chose to depart from the conventional scatterplot method typically used for this type of analysis and instead opted to explore this data’s visualization potential in map charts. I aimed to offer a clearer comparison of each country based on their respective data. To execute this visualization, I first needed to merge the data files in Tableau. While the air quality dataset included longitude and latitude coordinates, the car sales dataset only contained country names. Consequently, I performed the join based on the commonality of country names in both datasets. Then, I created a map using the air quality dataset as the foundation and then overlaid the car sales data for comparative analysis.

When cultivating the visualization (Figure 1), I felt that a color spectrum would be most effective. My first approach was to create a diverging color scheme where countries with poorer average air quality would be depicted in shades of orange, reminiscent of the smoky skies I witnessed in New York City. Conversely, countries with better average air quality would be represented in shades of blue, evoking the imagery often associated with environmental campaigns like “Earth Day.” Next, I added the car sales data to the chart as dots. The dot size would signify the amount of car sales each country made on average. 

I immediately realized that using dot size for comparison in this context would pose challenges in discerning subtle differences between countries. Alternatively. enlarging the dot sizes to enhance visibility would risk overshadowing smaller countries on the map with larger dots, potentially skewing the comparative analysis. Employing color to denote differences may have been a viable solution to this, however, I recognized the risk of overwhelming the viewer with excessive visual stimulation. Additionally, the base map’s use of five distinct colors spanning the spectrum, each with various shades in between, necessitated users to toggle between the legend and the map to decipher the significance of each nuanced color variation. This process prolonged the analysis and potentially obscured key insights.

For my final visualization (Figure 2), I opted for a sequential coloring approach, departing from the divergent color scheme used previously. Additionally, instead of representing the number of car sales with dot size, I chose to use a color spectrum specific to car sales data. I aimed to depict countries with higher car output in murky hues resembling smog while maintaining green tones for countries with better air quality. One challenge I encountered was calibrating the color spectrum for car sales to accurately reflect their impact on a country. I selected red to signify car sales, as it symbolizes caution, and the combination of red and green produces the smoggy brown color I sought.

Reflecting on the visualization, I find that the visualization effectively conveys my intended narrative, although it may pose challenges for others to interpret. Due to this, I wish I could create a legend that shows the blending of red and green to better represent the colors displayed on the map. The interactive feature of the map, allows viewers to better discern the car sales percentage by highlighting the corresponding red color when clicking on a country(Figure 3). Additionally, the predominant green hue on the overall map intentionally overshadows the hints of red in the graph, I achieved this by adjusting the opacity of the red to be lower than that of green. I did this by decreasing the opacity of the red the be less than green. In the future I would love to gather feedback through user testing to see if the message is conveyed clearly and to explore potential enhancements for a map chart visualizing data typically presented as a scatter plot.