The story of Sunshine PART 1
It all started 3 years ago…
It was a beautiful Saturday morning and we were heading out to a great windsurfing spot after looking up the weather on several apps — only to be disappointed once we got there. This happened often. Weather forecasts were either inaccurate or plain wrong. In the end, we’d just call up friends to check for other spots.
We worked as engineers and environmental designers by day and started tinkering with the idea of crowdsourcing weather reports by night. Our hobby soon turned into a full-fledged obsession. Then one day, we thought to ourselves: We can solve this problem. And so we quit our jobs to dedicate ourselves to it.
We hunkered down and built an app for beachgoers to post real-time weather reports, complete with photos and videos so that others could see what the conditions were like. Our users loved it and a tight community quickly formed. However, it was a niche audience and our technology wasn’t enough to disrupt weather forecasting. Beachgoers aren’t the only ones struggling with inaccurate, clinical weather predictions. It’s a universal problem faced by millions of people every single day.
So in September 2014, when Apple announced the barometric sensor on the iPhone, we pounced on it. This was the future of weather forecasting.
Redefining the Industry
Part of the reason weather forecasts are inaccurate and imprecise is the underlying technology. Traditional weather services utilize satellites, expensive weather stations and meteorological radars that are few and far between.
Expensive, decades-old technology dominates the weather forecasting industry
This technology often results in misses and these misses have real consequences for millions of people. We spent a year exploring the mobile sensors, building and iterating on forecasting models to make sense of the data.
Our initial results were incredible. What we discovered was that smartphones can integrate all sorts of environmental sensors and collect all kinds of data about their surroundings. In fact, they can actually map out the entire atmosphere and provide us with the data that can accurately predict the weather on a scale that’s never been possible until now. This was huge and yet we knew we had to overcome another challenge. We had the technology in place. How do we actually make people care? Dumping better data onto a screen didn’t seem like the right solution.
So what does a weather app need? Not tons of numbers, not technical charts, not stock photos or heat maps. We kept going back to the drawing board and ultimately came up with this list:
#1. Personal daily summary. A weather app needs to tell you if you’re going to feel cold or warm, if there’s going to be rain or hazardous weather later in the day, based on your comfort zone and your whereabouts (down to the street level!). Only then can you plan ahead and keep yourself safe and comfortable throughout your day.
#2. A breathtakingly immersive experience. What if we could create an environment that people could explore and immerse themselves in? A place where they could, with one swipe, experience what it’s like to be in Midtown NYC on a brisk October morning — or how the clouds descend upon Dallas when a thunderstorm is looming?
#3. 3x more accuracy and counting. As the Sunshine network grows, so does its accuracy. What if we could train our algorithms to be even smarter every day through the constant influx of data and our users’ live reports? In our beta test with just 200 smartphones in the Bay Area, our forecasts have been 36% more accurate than those from weather incumbents.