The Weatherman Wants Your Smartphone Data: Creating Hyperlocal Weather Reports through Crowdsourcing
The weatherman has competition, and it’s from a phone app. WeatherSignal, an Android app developed by Dutch scientists, tells the weather by crowdsourcing temperature readings from the sensors that keep smartphone batteries from overheating. Since smartphones are everywhere, WeatherSignal might one day let meteorologists make hyperlocal weather reports without building weather stations on every roof.
An individual smartphone is a crappy forecaster – people keep their phones in air-conditioned buildings and warm pockets – but the crowdsourced data from WeatherSignal’s 700,000 active users tells a much fuller story. In major cities, the app’s temperature readings are usually within a degree and a half of the actual temperature. Plus, WeatherSignal offers a range of readings traditional weather reports don’t even try to match: humidity, light intensity, air pressure, and even magnetic fluctuations. If you don’t mind clunky graphs, you can use the app’s data visualizations for a closer look at weather fluctuations.
WeatherSignal’s authors readily admit the app is a work in progress (it seems academics are more transparent than entrepreneurs). “This app was built in a tea-and-biscuits fueled two-week frenzy so there’s still a lot to be done,” they wrote in the FAQ section of their app. Users are invited to send in bug reports or manually input weather.
As mind-blowing as the technology seems, what WeatherSignal doesn’t do is predict the weather. Being able to accurately describe weather on such a local scale is an awesome scientific feat, but I do like knowing when I’ll need an umbrella. In its present form, the app is definitely better for gawking at cool data than for deciding when the weather will be perfect for a picnic.
Still, the app’s geologist authors are dreaming big. They recently published an academic paper speculating that if the number of smartphone temperature readings increase, their weather forecasts will eventually prove more useful than those taken from traditional meteorological stations. Densely packed temperature readings pulled from cell phone batteries might one day guide a kind of precision agriculture where highly local weather readings drive irrigation decisions. They’re also considering whether their data might be useful for scientists studying the urban heat island effect which makes cities a few degrees warmer than their surrounding areas.
Have you used WeatherSignal or a different weather crowdsourcing app? How does your app fare in the accurate weather reports department?