Visualising migration patterns

This wonderful CIRCOS-style chart (click to zoom) visualises migration data collected by the UN. Particularly noticeable are the flows from Mexico to the United States; and from India, Pakistan, and Bangladesh to the United Arab Emirates.

The chart is from Nikola Sander, Guy Abel, and Ramon Bauer. There is also an interactive visualisation of changes over time. Excellent data visualisation!

Palettes from paintings

I have been experimenting recently with R, my favourite statistical toolkit. The plot above (of the gender balance in Australian states and territories) uses a palette extracted from Vermeer’s famous Girl with a Pearl Earring (below) – a brown-and-blue diverging scale with some pink highlights. Interestingly, this is a palette which still works well under simulated colour-blindness. It looks good too.

Oh, and those numbers on the map? Partly related to an excess of males in the mining industry, and partly to an excess of females among retired people.

Twitter and global mobility patterns

A fascinating recent paper on, entitled “Geo-located Twitter as the proxy for global mobility patterns” (also reported on the MIT technology review) uses Twitter to study human movement (the study is based on a dataset of almost a billion tweets). The CIRCOS image below shows the top 30 country-to-country visitor flows, as estimated by the authors. Ribbon colours indicate trip destination, so Mexico-based Twitterers visiting the US are a major category. While the US is the most common travel destination, Russia is the most common point of origin.

There’s lots more in the paper: it’s well worth a read. Twitterers may not be totally representative of the world population, but there are still many interesting conclusions to be drawn here, and an opportunity for even more interesting follow-up work.

Network diagram from Hawelka, Sitko, Beinat, Sobolevsky, Kazakopoulos, and Ratti: “Geo-located Twitter as the proxy for global mobility patterns”

A-twitter with anger and joy

A recent paper from China studies traffic on Weibo (the Chinese version of Twitter), and finds that “users influence each other emotionally… the correlation of anger among users is significantly higher than that of joy, which indicates that angry emotion could spread more quickly and broadly in the network.”

The image below (from the paper) shows some of the emotional connections (red indicates anger, green joy, blue sadness, and black disgust). It would certainly be interesting to repeat this fascinating study in other countries!