Forecast

[forecastio-blog]

Skycons: Unobtrustive Animated Weather Icons

var v=new Skycons();v.add("sun",Skycons.CLEAR_DAY);v.add("rain",Skycons.RAIN);v.play();

Last year, Adam Whitcroft came out with his lovely Climacons: a comprehensive set of tastefully understated weather icons. You've probably seen them at one point or another, because virtually every weather app that has come out since then has been using them. (And for good reason: they're gorgeous!)

For most of it's development, Forecast also used them: they fit well with the aesthetic, and they were an easy way to bootstrap our design. They always seemed to be missing a certain je ne sais quoi, though, and we had a very difficult time identifying what, exactly, it was. Eventually, it dawned on us that, given the animations we had elsewhere in the app, the Climacons simply felt too flat, too static; we therefore set about making our own set of animated weather icons that felt more alive—but not so much so that they distract—which are the icons you now see on Forecast.

We are calling them Skycons, and they are now open source on GitHub.

Some Technical Highlights

Skycons are animated in HTML5 Canvases using JavaScript. In browsers that support it, we use the recent RequestAnimationFrame API to animate them efficiently.

Skycons are stateless: that is, the only information that is needed to render a frame is the current system time. It's somewhat tricky to develop this way, but the benefit is that the animations degrade gracefully regardless of the framerate they are able to be displayed at, and we don't have any data structures to maintain from frame-to-frame: this actually makes the animations lighter-weight, and hopefully less taxing on older or weaker devices, such as smartphones.

We intentionally designed the more-interesting weather conditions (such as "rainy" or "windy") to have much more motion in them than less-interesting weather conditions (such as "sunny" or "cloudy"), in order to naturally draw the eye to them.

There is currently an icon for every possible "icon" value that the Forecast Data API can return. As values are added to the API, you can expect us to make new Skycons, as well.

What's Next?

We're pretty happy with how Skycons came out, but there's always room for experimentation and improvement.

The code for Skycons is rather sloppy at the moment, due to its rapid development. It could really use some refactoring!

We developed Skycons in Canvas, because that's what we know and it was quick for us, but we were quite interested in experimenting with animated SVG images. This would make the Skycons declarative and completely self-contained, hopefully making them better-suited to a wider range of uses.

The set of icons was adequate for our uses, but there are quite a few weather conditions that aren't covered: thunderstorms, hail, tornadoes, tsunamis, etc. Adding these icons would make the package more useful as a whole.

Allowing dynamic customization of the animations, such as the angle of precipitation, could allow for any number of playful tricks, such as having the rain always fall towards the ground on your iPhone, using it's internal accelerometer.

As we get more time for playing around, perhaps you'll see some of these things pop up. Additionally, since Skycons are open-source, if you feel a burning desire to experiment with them, we invite you to do so! We're making the source code available in the public domain, with no restrictions whatsoever on its use, so use them however you see fit. (Naturally, we'd appreciate being credited for our work and for any interesting contributions to also be made open source, but we don't believe that either should be a legal requirement.)

So that's Skycons! We hope you enjoy them and find them a useful tool in the weather design toolbox.

[forecastio-blog]

The Forecast Data API

In addition to announcing the Forecast weather service, we would like to announce the launch of the Forecast Data API, a low-drama, general-purpose weather API for developers.

The Forecast Data API grew out of the original weather data API we built for Dark Sky. The Dark Sky API was designed to be very specialized for hyperlocal forecasting, and therefore was very tightly coupled to its data sources. As we tried to expand its scope, we found that cramming other data sources into it was rather like fitting a square peg into a round hole; in the end, we decided to rebuild the API from scratch.

As I write this, we have sixteen different data sources, covering the entire globe. To give you a sense of what the quality of our data is, here is a color-coded map:

Areas in black are covered by our data sources, but suffer from a lack of quality control: we cannot reliably check the results of our predictions in these areas, and so cannot verify their accuracy. (Expect them to be more accurate the closer they are to an area of a different color.)

Areas in blue are covered by our data sources and are also quality controlled by checking them retroactively against weather observations. However, the data coverage is low-resolution, and so may not accurately convey local weather conditions.

Areas in red are covered by high-resolution, quality-controlled data sources, which should provide accurate hourly forecasts down to the local level.

Areas in green are covered by Dark Sky's hyperlocal prediction system, and can provide down-to-the-minute forecasts for an exact location.

Notably, we can provide hyperlocal forecasts for the United States, United Kingdom, Ireland, and small parts of Canada. Hyperlocal coverage of Canada, Australia, and more of Europe are in the works.

Additionally, with so many data sources, we can check them against each other; where they agree, we can be more confident in our predictions. If you've tried the Dark Sky API and found its accuracy lacking, we invite you to give the Forecast Data API another shot: it should be far more accurate than its predecessor.

Now that we've developed a general-purpose weather API, we're trying to compete with the other weather APIs available around the Internet. We've found those APIs to be difficult and clunky to use, so we've tried to make our API as streamlined as possible: you can sign up for a developer account without needing a credit card, and start making requests right away--you can worry about payment information when your app is ready. Additionally, we've lowered our prices so that we're competitive with the other data providers out there.

If you're interested in reading more details about working with our new API, we invite you to check out our API documentation. If you want to start playing around, you can sign up for an account on our developer portal.

Finally, stay tuned: we'll be posting here in the near future about some of the neat technical tidbits we've come across while developing this entire system; if you're a developer, we're sure you'll find it interesting! You can keep on top of what we're up to through this blog or through Twitter.

[forecastio-blog]

Announcing Forecast

We’re extremely excited to announce the launch of Forecast, a new global weather service.

About a year ago, we released a little app for the iPhone and iPad called Dark Sky, attempting to do something new and interesting for weather forecasting, a field we think had become pretty stagnant. Approaching 100k sales, it’s been fairly successful; however, we’ve been continually asked for more: international support, longer-term forecasting, an Android app, and so on.

Rather than cram these things into Dark Sky, we decided to do something grander: create our own full-featured weather service from scratch, complete with 7-day forecasts that cover the whole world, beautiful weather visualizations, and a time machine for exploring the weather in the past and far future. You can access it from all of your devices, whether it be your laptop, iPhone, Android phone, or tablet.

On top of all that, we’re providing this data to other developers, in the hopes that a truly independent weather community can thrive in the era of increasing corporate consolidation.

A quick tour

How is Forecast different from other weather services? Here are a few features we’re particularly excited about:

Stunning maps and rain animations

We’ve taken our smooth radar animations from Dark Sky, and expanded them to allow you to scrub up to a week into the future to see how storm systems will develop and evolve world-wide. It is handy when trying to put a simple forecast such as “Rain next Wednesday” into context. (Try dragging the pin around to explore different locations on the map).

Down to the minute forecasts, for your exact location

We seamlessly weave the minute-by-minute forecasts from Dark Sky into this new service. We can predict when it will rain or snow — down to the minute — at your exact location anywhere we have radar coverage (currently the United States, the UK, Ireland, parts of Canada, and Puerto Rico, with more countries coming soon).

Our hourly and daily forecasts cover the entire globe.

The time machine

We’ve gathered hour-by-hour observations from tens of thousands of ground stations world-wide, in some places going back a hundred years. We expose it as a sort of “time machine” that lets you explore the past weather at any given location. We’ve also used the data to develop statistical forecasts for any day in the future. For example, say you have an outdoor family reunion in 6 months: with the time machine, you can see what the likely temperature and precipitation will be at the exact day and hour.

More intelligent text summaries

We also use this statistical data to generate more intelligent text summaries of the current and forecasted weather. For example: most people don’t really care about things like wind speed and humidity unless it’s exceptional for their location. Using the historical data, we know what “exceptional” is for any time and place, and roll it up into the summaries.

Data only when you need it

Existing weather apps tend to fall into two categories: Weather Glut, which target weather geeks by cramming in lots of data most of us don’t care about, and Weather Fluff, which attempt to reduce the interface to a bare minimum for the sake of simplicity.

Where do we draw the line? You may never care about the wind direction  —  until the one time you go sailing. You may never look at an hour-by-hour breakdown of the upcoming day  —  until a hurricane is about to hit.

With Forecast, we’ve tried to give you the best of both worlds by keeping the interface as simple as possible, but also providing links to additional detailed information, to let you delve deeper whenever the need or mood arises.

World-wide support

Try checking out the rainy season in Bogotá, Colombia, or see how Santa’s faring at the North Pole!

It works on all your devices

There’s a web interface for your laptop and a mobile version for your iPhone, Android phone, or tablet. You can install Forecast directly on your phone simply by visiting forecast.io on your device and adding it to your home screen. It’s completely free.

An API for developers

Under the hood, Forecast is powered by a data API that we are making available to others. The API makes it extremely easy for developers to retrieve weather data for any location on Earth. More information can be found at our Developer site.

If you make a weather app, we’d love it if you’d consider using our data. Don’t hesitate to contact us with questions or suggestions!

What’s next?

This is just the beginning. Over the coming months, we hope to continuously improve the service and expand our coverage. We’ll also be adding user accounts that’ll give you access to notifications and the ability to sync your saved locations across all your devices.

We’ll also being regularly posting updates on this blog, exploring in detail the various components of both the Forecast website and our upgraded data API.

Finally, what about Dark Sky? Don’t worry, we’re not abandoning it — in fact, we’ll have an update in the App Store in the coming weeks that will add in our improved weather forecasts and add support for the UK.