A team of London-based smartphone app developers has combined signals from temperature sensors in smartphone batteries to accurately forecast weather in major cities.
Using the OpenSignal Android app, the team has created a crowd-sourced weather forecast platform that collects data from hundreds of thousands of smartphones. The built-in sensors that have made the venture possible are originally designed to prevent smartphone batteries from overheating. However, they have proven to be also a particularly accurate source of up-to-date meteorological data.
Using the measurements of the smartphone micro-thermometers, the team has managed to predict weather in eight major cities around the world with a 1.5 ℃ deviation. The team believes that if more users join the system, the accuracy will further improve, filtering out other intervening factors more efficiently – such as whether the phone is currently being used indoors or outdoors or whether the user is playing a game on it.
The results of the study were published in Geographical Research Letters earlier this week.
“The ultimate end is to be able to do things we’ve never been able to do before in meteorology and give really short-term and localised predictions,” said James Robinson, co-founder of OpenSignal. “In London you can go from bright and sunny to cloudy in just a matter of minutes. We’d hope someone would be able to decide when to leave their office to get the best weather for their lunch break.”
The technique could provide a valuable complement to conventional weather forecast that relies on meteorological stations. Despite these stations being equipped with accurate and advanced sensors, their network only has a rather limited density, which in turn only allows for limited spacial resolution of the predictions.
On the other hand, as there are millions of smartphones distributed among the public all around the globe, the smartphone generated data could help forecast even the tiniest differences in temperature down to a level of a city block.
The OpenSignal app collects data only from users who have joined the network voluntarily. To date, 700,000 individual users have signed up with about 90 per cent of them having agreed to provide statistical data, enabling the OpenWather app to analyse over half a million temperature measurements a day.
The discovery that smartphones could actually predict weather was in fact a by-product of research looking into differences in temperatures between smartphones using 4G networks and those relying on the older types.
“Just for fun we started looking to see if there was a correlation with anything else,” said Robinson. “We got some London weather data for comparison and found the two sets of temperatures were offset, but they had the same sort of shape.”
After London, the team has achieved similar results in other cities with substantially different climates – in Los Angeles, Paris, Mexico City, Moscow, Rome, San Paulo and Buenos Aires.
“It was amazing how easily the correlation sort of popped out,” said Robinson. “We didn’t do any handpicking of data, it just emerged.”
Despite not aiming at competing with conventional weather stations, the team believes that once smartphones become equipped with dedicated sensors to monitor the environment, the smartphone-based weather forecast could eventually take the upper hand.