The New Weather Era: Apple’s Data Independence Journey

Written by Steve Gifford

June 10, 2022

Apple announced WeatherKit at this year’s WWDC. It looks a lot like Dark Sky’s old forecast API, which is not a surprise since Apple bought the company behind Dark Sky. It took over two years to get that out the door, but I want to discuss something else. Did you notice Apple has largely ditched IBM’s Weather services?Are they interested in data independence?

That’s the part I find most interesting for us. But first, some disclaimers.

Dark Sky Acquisition

I have yet to learn what the internal deliberations at Apple were like, nor do I know what the terms of the Dark Sky deal were. My old consulting company used to work for the Dark Sky folks, but they went dark as soon as they were bought. I have not talked to them since, which is standard for Apple. So, all this is speculation.

I can say this. Apple appears to no longer use much IBM data for its weather-related products. They seem to be moving that all internally, and that’s fascinating.

Weather Data Sourcing

We get most weather data for free in the U.S. NOAA provides a bounty of real-time sensor data that companies, big and small, process into useful information for their consumers. In other countries, it gets more complicated.

The big companies often do post-processing on that to tweak and adjust data sets based on past patterns or very up-to-date real-time data. That’s a good thing, but they lean into that for sales.

Managing the flow of weather data is tricky, and it’s worth something. Running your own blended statistical model is worth something, and those newer ML models are worth something. But much like cable in the 90s, it would be nice to unbundle it.

Weather Data Independence

Again, I can only guess from Apple’s public actions. It looked like they wanted more data with fewer restrictions. They seem to have wanted to do more with that data as well, and they wanted access to it in a way IBM found uncomfortable (or expensive). You have to wonder if user privacy was mixed in there, too.

That’s what the big weather companies are like. They hold their data tight and you can only view it through a tiny window. I’ve had this problem with clients trying to visualize these data sets. They just don’t have access.

Our Weather Solutions

This is our blog, so of course, it’s about us. We focus on weather/sensor visualization—the kind of stuff they do well in weather apps because we wrote that for several of them. But now we’re doing it for you, the enterprise customer who can write decent-sized checks.

What does it have to do with Apple? Nothing directly; it’s an analogy. Just as they broke free of their big provider to do more interesting things with the data, so can you. How about:

  • Mixing in your one proprietary data set you have
  • Much better data visualization for mobile and web that you can use in anything you like
  • A forecast endpoint that’s all yours with predictable costs
  • Access to interesting new models from scrappy third-party forecasters
  • Complete privacy. We won’t mine your queries for anything
  • As much access to the data as you’re allowed by the provider

We’re building that. Some will be ready sooner, some later, but that’s the plan.

If you’re looking to break free of your existing provider’s restrictions or contemplating spinning it up yourself, contact us