I’ve always wondered this. Turns out, no. Google Maps optimizes for the current user only. It blends real time data collected from other users with historical data. But doing this for each individual at scale can help to balance traffic as a by-product.

I ended up reading World scale inverse reinforcement learning in Google Maps, which I found intellectually humbling. Its fun to realize how deep things go when you look closely.

Returning to my DuckDuckGo search results, I saw this image of global tweet density in a heat map. This is Twitter only, so not terribly useful — China is completely dark. But it reminds you how connected we all are, and reveals in which countries Twitter is popular.

That image gave me a hankering for a proper global Internet heat map, and this one does the job. It was built by pinging every possible IP address and then geopositioning any hits. This somehow only took 5 hours to run.

Seeing this made me wonder how it would compare to a population density heat map. World Population Density is OK, but it’s too light to really see. Google’s Global Population Explorer has better contrast, but the country outlines make it hard to look at zoomed out.

Africa is the outlier when comparing population density with Internet density. It has pretty substantial population density over large areas that show relatively little Internet activity on the heat map.

This map is interesting to zoom in on. For example, I never realized how many people live in northern India, just below the Himalayas. And Russia’s population is a huge hub and spoke centered on Moscow.