I chose HIV as my project topic based on the article, HIV's Southern Trap in The New York Times. While I had heard and read about HIV and AIDs in the ’80s, it was when I went to undergrad (and outside of my suburban bubble) to study photography and film in Philadelphia where I was made more aware of HIV and AIDS. HIV’s Southern Trap awakened me to the startling reality for many men of color. I felt compelled to understand how HIV had changed in twenty years.
This was one of many questions I had as I read HIV's Southern Trap.
In exploring these questions, more questions came to the surface and I used these questions to build and sketch potential outlines for a narrative. Writing words, phrases, questions, and ideas on paper helped me to see threads and possible connections. Would the data help to answer my questions? Would data disprove some of my assumptions? Would the data support the NYTimes story?
Initially I had plans to present how rates of HIV changed over twenty years — 1998 to 2018. But I soon discovered the data wasn’t nice and clean or even consistent. I downloaded twenty years of PDFs from the CDC, extracting the data using Tabula and while converting numbers to rates per 100,000 people, I discovered the data in 2007 was rather odd.
I'm not an expert on HIV data reporting so I don't know exactly why data in 2007 created a visual dip, but my guess was that in 2008, HIV reporting to the CDC became standardized. Based on this quirk, I decided to modfiy the direction of the visualization to show approximately 10 years of change from 2008-2018.
Caption: Covers from the CDC’s HIV Surveillance Reports; Census Bureau population data to convert total numbers of HIV diagnosis to rates per 100,000; CDC Atlas Plus data portal; Tabula software icon; Sample MS Excel sheet for conversions and notes about the data.
Designing my first information design taught me to explore data by experimenting with different types of visualizations. Tools such as Flourish, RawGraphs and Adobe Illustrator allow me to quickly see patterns, outliers, missing data and gives me a sense of whether the encoding is the best or could even be pushed further.
Because there is a lot of data about HIV, I used this exploration stage to determine what threads, if any existed between them. How does each visualization present something new? Is it clear? How does the size and shape impact the overall structure, flow and narrative? What assumptions did I make? What edits need to be made? What relationships exist between them? What is missing?
Critique was essential. Each iteration incorporated feedback from classmates and Alberto.
Data can be messy. Accuracy and transparency is critical so, in an ideal world, I would have worked on this visualization with a team experts who study the epidemiology of HIV everyday.
Complexity requires focus. Perhaps this is typical of beginners but the more I learned about HIV through data and through primary and secondary literature, I went down rabbit holes that were fascinating but not needed for this graphic.
Your “clippings file” can surprise you. The heatmap was inspired by a visualization about U.S. Unemployment I had clipped from the Data Illustrator gallery while completing another assignment. While searching for a how-to tutorial in my notes, I stumbled upon the U.S. Unemployment heatmap (originally created by The Washington Post) and thought it might be an interesting and different way to show how rates of HIV have changed by race, age and gender between 2008-2018.
Make the time to learn. Data Illustrator is not intuitive so, while I was excited by the prospect of presenting rates of HIV in a more compelling way, I was daunted by having to use a burdensome tool. In the end, the extra hours and time I put into the lesson and the countless trial and errors were worth it.
Epidemiology is fascinating. Over the last several years, my interest in health has evolved from wanting to cook and advise people about nutrition to understanding the causes of health outcomes. For me, it is a new way to look at the world and see possible connections between policy, behavior, geography, economics, and more.
Caption: A draft of line charts showing the US rate of HIV compared to each state plus D.C. between 2008 and 2018