If there's one thing your phone's app store might have too much of, it's weather apps. My search turned up more than 7,800 tools, including the Weather Channel and Yahoo Weather, all of which do basically the same thing: relay forecasts from remote weather stations across the country.
But maddeningly, despite this plethora of options, there are still times when we leave the house without an umbrella only to get caught in a freak rainstorm that our weather apps failed to predict. That's where a new smart app called Sunshine comes in. Sunshine wants to make weather forecasting more accurate and more personal by relying less on remote weather stations and more on real-time reporting from users.
The key to Sunshine lies inside your smartphone itself: the barometer. Every new iPhone 6 or 6 Plus comes with a built-in sensor that measures atmospheric pressure, a key indicator of weather changes. This means thousands of little weather stations are already scattered across any given city, tucked away in purses and pants pockets. Sunshine taps into these barometers to create hyper-local forecasts. "It's called distributed computing," says Katerina Stroponiati, founder of Sunshine. "All the devices are connected so we can provide real-time info when you need it." Traditional forecast tools only update every six hours, according to the Sunshine website.
Other apps, like PressureNet and WeatherSignal, already use crowdsourced data. But Sunshine has something else to offer: It's a smart weather app that learns, meaning it gets to know you over time. "Everybody else is doing the same thing," Stroponiati says. "They just provide you numbers, but they're missing the personalized experience. They're not related to your routine, your schedule, or how you feel." When you open the Sunshine app you are asked if you would rate the current temperature as "cold." Using your response as a reference point, the app begins to accumulate information about your personal comfort preferences and sends alerts or recommendations to match those preferences.
For example, say you have a pollen allergy or sensitive skin. Sunshine could alert you on days when there's a high pollen count or the sun is particularly strong. The app also learns your commute and, if it's raining at your office but not at home, it will recommend you bring an umbrella. "We try to keep you in your comfort zone based on the feedback you provide us," Stroponiati says. She compares Sunshine to the revolutionary Nest Learning Thermostat. "Before Nest, nobody would care about the thermostat," she says. "It's totally boring, we just have to use it. But Nest came and gave a more personalized experience and now we are proud of our thermostats."
Sunshine collects data in the background while you're out and about, using minimal battery power. While you don't have to input any data manually for the app to work, its creators hope you will. For example the user can report a hazard like an icy road, or sudden fog, or a particularly windy street.
The app's accuracy will depend almost entirely on having a well-distributed user base. The minimum for a good forecast is three phones per mile, Stroponiati says. Currently the app is in beta in the San Francisco area with plans for a full launch sometime in April.