For this lab, I continued to work with world migration data. Since both of my previous lab projects also focused on the same subject matter, and also featured maps, I chose to focus on a particular aspect of the data set that I had not previously explored. In addition to total migration figures to and from each country, the Bilateral Migration Matrix from the World Bank broke down migration by gender, and I was interested in seeing whether certain countries had more emigration from either gender. Although the data itself was bucketed into five 10-yr buckets (1960, 1970, 1980, 1990, 2000), I chose to show the cumulative migration for each country over the entire 50-year span. This was partially due to constraints within the platform (showing changes over time would have necessitated creating five separate maps), but also because I was interested in seeing what the overall migration pattern for a given gender or country was–which might demonstrate a much larger and more profound trend.
In looking for existing visualizations on this topic, I was interested to find that cumulative migration by gender was not represented. I found several sources that examined the same data by decade, but non provided visualizations for the total numbers.
This map which shows then number of women per 100 men in the U.K. population achieved good visual comprehension by using a similar system (ratio represented by color) that I hoped to show with my viz.
This map showing sex ratio (not migration) was at first very interesting to me in that it showed at first glance a much higher number of males in the world population than females. Upon closer inspection, the legend explains that the color assignment is reverse to popular assumption; this was utter insanity to me, and demonstrated the pitfalls of counterintuitive color use.
This world map, which uses color much more appropriately, also offers some insight on my final visualization, showing that some countries with higher female populations also have higher levels of female migration from that country, presumably because their marriage opportunities are few domestically (see: Russia).
This data set sourced from the World Bank, the Global Bilateral Migration Matrix, already included gender of migrants, and also included a total. I filtered the data to only show male and female; the total was not necessary for this visualization, and could also be summed within CartoDB. After filtering, I had to sum all rows for a given country in order to get the cumulative migration for each gender by collapsing the time variable. To do this, I created a Pivot Table in Excel and set the value function to SUM so that all values would be added across each country’s five entry rows.
To best represent the gendered composition of each country’s total number of migrants, I decided to calculate a ratio for each country. This would put each country’s number female and male migrants on the same scale. This ratio was created by a formula function in Excel which divided the number of female migrants by the number of male; this new column ranged in value from approximately .4, indicating that more than twice as many males have migrated than females, to 2.34, indicating the exact opposite.
View map on CartoDB by clicking on image.
Creating the Map
Using CartoDB, which contains existing shapefiles for countries of the world, I imported the spreadsheet with gender ratio data. I joined the shapefile and data by the country column.
I chose to use the most obvious, if somewhat overused and stereotypical colors to represent levels of male and female migration (on blue to red spectrum, with cream as a neutral color in the center) in order to make the map instantly comprehensible to viewers. However, the initial assignment of colors to values was misrepresentative of the data, as the data is not normally distributed (negatively skewed) and I had to assign custom thresholds for each shade. This shifted the map from being more or less evenly split between blue and red, to being predominantly pink and red. In the final version of the map, any country whose female to male migration ratio falls between 0.9 and 1.1, indicating a net neutral migration of genders, is colored cream.
Once the colors were set, using a range of hex codes from Pantone to standardize them, the visualization made it clear that the majority of countries have far more female than male migrants leaving their borders, so I chose to title the map “Girls on the move.” I ran some simple statistics on the ratio data in order to provide some simple facts that would contextualize the map (such as the fact that 80% of countries have more females than males leaving, and that 53% of total world migrants are female) and added these facts in a text box below the title.
Discussion and Future Work
As already mentioned, this visualization strongly shows that more females than males have migrated from 80% of world countries over the past 50 years. Although the source data does not specifically identify how migrants are identified and what their age ranges are, since the World Bank compiles this matrix from census data, it can be assumed that this includes all migrants such as adoptees, refugees, workers, etc.
It is therefore reasonable to posit that some of these countries, such as China and Russia, which are popular countries for international adoption, can attribute some portion of their female emigrants to adoption.
The countries with the very deepest red, those with female to male migration ratios above 2.0, such as Zambia and Zimbabwe, may indicate regions where women have few rights and migrate for better opportunities, both economic and social.
In the context of a mostly pink and red map, those blue countries with higher male migration numbers stand out rather starkly; it’s interesting to think about the reasons for these patterns, which may include economic and work opportunities. For countries like Mexico, this is certainly plausible; this is also true for the blue countries in Northern Africa and the Middle East which are and have been subject to frequent wars, unrest, and economic depression.
It is almost important to note that since this data is based on official censuses, these figures likely do not include ‘illegal,’ or undocumented migrants. It’s entirely possible that including such data would paint a very different picture. Such migrants might include those who voluntarily migrate for economic opportunity, but also those who are involuntarily migrated–such as girls and women subject to human trafficking and the international sex trade.
Lastly, this data of course adheres to a gender binary and is therefore not representative of gender-non-conforming individuals; future data collection would need to include more options for gender identification to make this kind of representation possible.
Further work could also include examining this data on a time-series basis in order to identify the change in trends over time.