Twitter engineers Jimmy Lin and Miguel Rios have analyzed data from four different cities worldwide to show the seasonal variation in tweet patterns. Using data visualization to generate insights into human behavior, they found differences in activity and sleeping times.

They examined Tokyo, New York, Istanbul, and Sao Paulo, and found that Tokyo Twitter users have a relatively constant sleep schedule throughout the year, but those in the other cities exhibit seasonal variations. Also, Japanese users’ activities are concentrated in the evening, whereas in the other cities there is more usage during the day.

The gradient from white to yellow to red indicates the amount of activity (light to heavy). Each tile in the heatmap represents five minutes of a given day and colors are normalized by day.

Twitter Engineering

Photo by Rich