For this lab, I was interested in looking at migration patterns from countries across the world. I was originally interested in finding data about emigration from the United States to other countries but wasn’t able to at the time. Based on the net migration data that I found, however, I was interested in examining the countries and/or regions that had the highest levels of emigration or immigration. I thought that by further looking at the income levels of these countries and/or regions, I could make some guesses as to the reasons for migration.

Since this was geographical data, I knew I wanted to incorporate a map of the world. The primary information in this data set (net migration) also lends itself to some kind of heat map or color saturation, to show variation over time.

Inspiration

This world map from Wikipedia showing net migration in 2014 had elements of what I wanted to show in that it includes all countries and uses color to indicate positive, neutral, or negative migration. However, the use of color was not intuitive—while a diverging color to show positive (immigration) vs. negative (emigration) did make sense, adding a third and unrelated color to show neutral migration did not. This map relied heavily on the legend, which is not ideal.

 

This map, while on a different topic and showing only the United States as opposed to the world as was suitable for my data set, had elements of the time and color saturation I was hoping to use. Unlike the first map, which used three colors to indicate net migration, this viz used only one color and adjusted saturation to show variation in levels.

While very simple, this bar graph was effective at showing the highest and lowest levels of net migration. The fact that it is sorted from highest to lowest makes it impactful. The color choices also made a lot of sense to me here—green indicates a positive migration flow (immigration) and red indicates a negative one (emigration).

…

 

The net migration data used for my visualization came from the World Bank’s data portal; the original data set provided only net migration by country between 1964 and 2014. I supplemented these data by adding region and income level information for each country, using separate data sets and the V-lookup function in excel. The original data set was in table form and had to be normalized by transposing columns for 5-year ranges into rows. This was done using OpenRefine.

The visualizations were created using Tableau. The main viz, of the world map, was generated by putting latitude and longitude onto the worksheet, and then assigning the net migration variable to color selection. I chose a red-green diverging color palette, as inspired by the Index Mundi bar graph showing world migration, with red indicating a negative flow of migrants and green indicating a positive one. The center of the spectrum, indicating a neutral flow of migration, defaulted to grey, which seemed appropriate. I added the time variable to the page tool in Tableau, which made the map animated across the entire time span. I removed country labels to make the viz clean and simple, but left the legend so that viewers could see the range of migration flow across colors. I also included a search function so that users could target a specific country (and later applied this search function to work across the entire dashboard).

In addition to show migration by country and/or region, I was interested in showing trends by income level. Since I had already assigned net migration as the color variable on the world map, it seemed like the viz would be too cluttered to add additional levels of visual information for income level (although income level does show in the detail for the map when the user hovers over a specific country). A line graph seemed like the best way to show trends of migration by income level over time. This viz was slightly troubling to me—I initially tried to use a different color to indicate income level, but found that once the vizzes were combined in a single dashboard, the use of color to describe multiple variable was too confusing. I therefore switched the line graphs to grayscale and broke the single graph into small multiples with labels for income level.

Although the world map showed net migration over time for each of the 5-year buckets provided in the original data, it did not indicate cumulative net migration for the entire 50-year span. I decided to represent this information in a bar graph similar to the Index Mundi graph from the inspiration viz. While this produced a very long bar graph that would not easily fit on a dashboard, I think it tells a powerful story in the disparity of net migration levels, with the bar for the United States being several times longer than its nearest neighbor (Russian Federation). I used the same red-green diverging color assignment here to stay consistent with the world map.

Several things shown by these vizzes were not surprising. For one, it was not surprising to see that the United States has always had a consistently positive flow of immigration. It was also not surprising to see strong negative migration flows from certain countries during times of war (i.e. Afghanistan in the 1980s and 2000s; Syria in the 2010s). Interesting to note is that for each of these countries, the 5-yr periods of low net migration/high emigration, are followed by a 5-yr period of high net migration/positive immigration. This could suggest that refugees from war time are resettling in their homelands following periods of unrest and violence. Also unsurprisingly, net migration levels were highest for high income countries overall, indicating that more people immigrate to, rather than emigrate from, high income countries. Somewhat surprising was that lower-middle income countries show lower net migration than low income countries, which have a more stable net migration level. This could possibly be explained by the fact that people in the lowest income regions have fewer resources to migrate, whereas lower-middle income countries may have seen a decline in income level, inciting a desire to leave for higher-income regions.

In future work, I would like to analyze more specific migration flows, including looking into which specific countries people from another given region or country most frequently migrate to. If combined with research on cultural norms and other cultural data such as religion, political affiliation, gender, cultural beliefs, etc, this could lend insight into reasons for migration, and the ways in which people seek similar or different environments to live in. I intend to do at least the beginning stages of this work in creating a network visualization for this class later this summer.