The world around us has changed dramatically in the last 20 years, and with it how we work, play and communicate — think cloud computing, social networks, smartphones — yet weather forecasting has not. Traditional sources still include radar, weather stations, satellites, and that’s about it. These limited sources are the reason weather forecasts, even on nice-looking apps, are far from dependable (ever wish you’d known when it would start raining to avoid a day of wet shoes?)
But how many devices around us could sense the weather? We all carry in our pockets a device with more sensors than we know; IoT devices are everywhere, and they may have sensors that can be repurposed; the air around us is rich — filthy rich! — with wireless signals that behave differently depending on the medium in which they travel.
That’s where we come in. At Tomorrow.io we gather, reverse-engineer, and analyze this unique data. We sense the weather, and we do so purely by listening — really listening — to the world around us. Utilizing our strong data partnerships, we apply passive sensing techniques through millions and millions of data-points that others may find useless, but we extract gold from. The largest «cab» company has no cars. The largest «hotel» chain has no hotels. We are the best weather forecast provider, and we have no hardware sensors.
The New Technologies team at Tomorrow.io faces that impossible challenge — we locate new sources of weather and plug them into Tomorrow.io’s top-notch models. Where others have tens or hundreds of data points, we have thousands and millions.
Like all repurposers, it’s a matter of looking at the world differently. Where others see cell towers, TV satellite dishes, air planes, drones or cars, we see potential sensors. We often start with a crazy idea (could windshield wipers tell us something about weather?) and then put together a POC experiment to get to a «go/no-go» decision. Once we get a «go» and collect enough data to analyze, we leverage Tomorrow.io’s proprietary validation tools to make sure the data we’ve acquired can indeed increase our quality and granularity of measures. We then deliver our prototypes to the engineering team who face the difficult challenges of scalability and implementation, and move on to our next challenge.
These new data sources from the connected world, coupled with top-notch data processing algorithms and prediction models, result in MicroWeather, hyper accurate and specific weather forecasts — street-by-street and minute-by-minute, globally — providing a new level of accuracy and specificity in weather forecasting.
Kids are often taught about the 3R — reduce, reuse, recycle. It might be time for a fourth R — repurpose. Looking at the world in new ways and finding gold where others see nothing.
