Aart Overeem of Wageningen University, the Netherlands, and his colleagues exploited an existing Android smartphone app to collect and select around 1.3 million temperature readings from eight cities around the world, averaging more than 800 readings per city each day. The app gathered data from phone-battery temperature sensors and uploaded the readings when a data connection was available. A simple heat-transfer model accounted for the effects of body warmth and clothing, and allowed the researchers to estimate daily temperatures in urban environments. The approach could be used to collect data at a finer scale and lower cost than is currently possible with weather stations, the authors suggest.