Contraceptives vs ClassData Visualization
Contraceptive use, income, and education are all intimately linked, so to speak. This data visualization explores the relationship between such socioeconomic indicators on a global scale and in Canada.

Despite its status as a developed country, wealth disparities in Canada leave many without access to effective and affordable contraception.

Audience

Educated lay audience

Tools

Tableau Prep, Tableau Desktop, Excel, Adobe Illustrator

Client

Prof. Shehryar (Shay) Saharan (University of Toronto)

For a detailed description of the development of this piece, including references, please consult the dossier

Design & Development

In my first few attempts at visualizing this topic, I did not graph the data. I wanted to get a sense of how I might arrange elements to transform raw data into a cohesive narrative. I also focused on finding a graph type that would suit the story behind the data. The birth control packet provided a convenient form of bar chart to represent unmet need for contraceptives.

I tried out a few different visual elements, including a banner with repeating icons of common contraceptive methods. It was excluded from the final because it did not match the clean, minimalist look I was aiming for.

Data Processing
This piece posed a unique problem:

How do you effectively layer data when information is missing?

When data is incomplete, it is tempting to hide blank spaces that feel as though they undermine the credibility of the work. In reality, the absence of data is informative. It identifies gaps that require further inquiry.

Survey data on income and education, as measured by test scores, had previously been compiled by Our World in Data. I prepared it in the following order:

In Tableau Prep, I combined that income and education data with data from the UN on contraceptive use and unmet need. These datasets had some overlap, but many countries were excluded due to missing data on multiple variables. The first graph resulted from the intersection of countries with mostly complete data.

Missing data on unmet need was represented with empty circles. This circumvented the issue of excluding an even larger number of countries.

Countries with full data were filled with an image of a small, round pill resembling a typical birth control pill. This design element is a subtle tie-in with the birth control packet in the second figure.

Data on method of contraceptive use vs income and education was de-identified and unlinked in the Statistics Canada dataset. Since these individual-level associations were excluded (likely for privacy), the Sankey diagrams I had originally intended to use were not possible. Furthermore, the dataset had significant ranges of error for the reported numbers and I wanted to capture this variability in my visualization.

Multiple attempts at resolving this issue can be seen in the rough sketches here.

The final solution was a collection of small bar charts, highlighting overall differences in contraceptive use based on education vs income, with multiples for each contraceptive method (e.g. condoms, IUD, ...). This representation highlights the correlation between more effective, permanent methods of contraception and higher income. That conclusion points to the link between socioeconomic class and bodily autonomy.

The faded error bars represent the ambiguity posed by large confidence intervals without distracting from the narrative of the graphs. These graphs were created in Excel (left) and refined in Adobe Illustrator (right).

The final piece was originally in a vertical layout. This was an inconvenient format for web viewing and printing, although I felt it supported the narrative more effectively than a horizontal layout. One advantage of the horizontal layout is its direct comparison of global vs Canadian contraceptive use.