This article was published on June 25, 2015

How emoji and cherry blossoms helped fix weather reports


How emoji and cherry blossoms helped fix weather reports

There are 29,000 weather stations across the globe. 20,000 of them are in the US. This leads to huge gaps in coverage and means weather forecasters are often making educated guesses.

That’s a big problem for us all and the secret to fixing it begins with cherry blossoms. No, really, stick with me here. Japan’s obsession with spring flowers will make weather forecasts better for everyone.

In Japan, the cherry blossom season is a big deal. Hanami (‘flower viewing’) is an important tradition and there’s fierce competition to accurately forecast when the trees will start to bloom.

Screen Shot 2015-06-25 at 11.24.23
The 2015 forecast by the Japanese Weather Association

From 1955, the Japan Meteorological Agency, the country’s national weather service, provided forecast data on when the blossoms would flower across the islands. Five years ago, it stopped. The reason? Other organizations had better data. And that’s why it’s getting easier for you to know whether you’ll get sun, rain or snow, wherever you are.

The <3 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol' founder Boris, and some questionable AI art. It's free, every week, in your inbox. Sign up now!

The Japanese Weather Association, whose forecast is shown above, was the country’s first private weather company. It was founded in 1950 and employs 580 people. But it has been out-paced and outgrown by a much younger rival.

Weathernews is now the world’s largest private meteorological services company. It’s headquartered in Japan and has over 700 staff worldwide in 40 offices spread across 15 countries.

It was founded in 1986 by Hiro Ishibashi, whose interest in accurate weather data was partly inspired by the loss of a merchant ship during severe storms.

WeatherNews-sea
An illustration of the sea from Weathernews

Ishibashi’s son, Tomohiro Ishibashi, has followed his father into the family business. He told me how it managed to create a more accurate cherry blossom forecast by sourcing data through its mobile apps:

Japanese people are crazy about the cherry blossom. They party under the trees. But the problem was that the old forecasting model was not accurate, because of climate change, it doesn’t work anymore.

So we gathered data from our users. We said to them, ‘Please send a photos of the buds to us. As many as you can.’ We got images from across Japan and put all of that data into our system. It is much more effective.”

Screen Shot 2015-06-25 at 13.51.29
Dark Sky

Apps like Dark Sky have introduced innovative ways of displaying the weather. Ishibashi is a fan but says that’s not enough:

In the database layer there is no differentiation, only the user interface can be their differentiator. Apple, Google, Dark Sky…their styles are interesting but the data source is the same. We can provide totally different content compared to other weather apps.

He compares the process of improving forecasts to running a race: “It’s like the 100m sprint. Adjusting the algorithms just leads to a slight improvement over time. Using crowdsourced data is like suddenly running on a moving sidewalk.”

Meet the mob

Weathermob

In May, Weathernews acquired Weathermob (“Waze for weather”), a Boston-based social service. It had already launched Sunnycomb, its own app for iOS and Android (“the Instagram of weather”), using a similar approach in 2013. Both now coexist under their parent company feeding information into the same growing database.

Weathermob’s CEO – now Weathernews’ Chief Editorial Officer – Julia LeStage had a previous life in TV commissioning. At UK station Channel 4, she was responsible for the first series of Big Brother.

She’s now inspired by the mingling of information from smartphone users, traditional weather data and algorithms:

For a non-weather person, the beauty of a place like Weathernews is that they have meteorologists that are building the models that can receive this crowdsourced data. The merger of tech with the weather knowledge of legions of meteorologists who can receive that information is the game changer.

LeStage argues that the company has changed the way weather forecasts are seen in Japan. It’s certainly the dominant player there with well over 20 million downloads for its main app.

Ishibashi says the appeal of Weathermob was its community:

There’s a bunch of apps in the weather industry, but there are only a few that focus on community. We use reports from users, sky pictures and other data to make our forecasts much more accurate. Weathermob has a strong community of weather enthusiasts.

It also has an interesting approach to letting users report the weather around them, tapping into the ubiquity of emoji. LeStage explains:

It’s emoji led because as a former daytime TV producer, it never occurred to me to talk about something without putting emotion at the heart of it. About a third of our users are British and English is the lead language in our app, but we like the idea that weather is language agnostic.

But making it easier for people to share their own reports doesn’t solve the lack of weather stations providing accurate meteorological data. Weathernews has an answer for that.

Station to station

Screen Shot 2015-06-25 at 13.41.10
World weather station map by wws-map.com

The company has just partnered with Netatmo, a French company behind the largest network of privately owned and smartphone connected weather stations. It’s suddenly able to draw a much better picture of the world’s weather.

Ishibashi says:

Accurate weather forecasting is only as powerful as the data used to create the predictive models. By integrating Netatmo’s worldwide network with our own large and existing data including from Sunnycomb and Weathermob, we can provide more personalized, accurate and localized services.

Netatmo’s devices are spread across 175 countries. They’re made up of an indoor module that measures temperature, humidity, decibels and CO2 in your home and an outdoor one which tracks temperature, humidity and barometric pressure.

All of that information is now part of the overall data included in Weathernews’ model, meaning it can now make its forecasts hyperlocal. It’s as close as you can get to the accuracy of looking out of the window or even, god forbid, going outside.

LeStage paints a utopian vision: “Many different countries collecting data and sharing it together, but it could be world changing in the way we understand the weather.”

Ishibashi goes even further. He is intensely passionate about realizing what he sees as his father’s dream:

We believe weather information services help mankind…weather information contributes to the health of the people. I have always followed his dream. It is our mission to make it a reality.

It all reminds me of Samuel Johnson’s comment on the English and weather:

It is commonly observed, that when two Englishmen meet, their first talk is of the weather; they are in haste to tell each other, what each must already know, that it is hot or cold, bright or cloudy, windy or calm.

With our smartphones in our hands, we can tell not just those we meet but the world about the weather around us and in doing so, we might just be contributing to data that finally fixes forecasts forever.

Read nextCarrot Weather will be your favorite new sarcastic weather app

Featured image: Shutterstock/Vasin Lee

 

Get the TNW newsletter

Get the most important tech news in your inbox each week.