There was a time, long before internet and mobile devices existed, when you had to wait for the next day’s newspaper to get the weather. In 1982, The Weather Channel changed the landscape by bringing weather to cable television. That’s when we really started mapping out the atmosphere. A lot has changed since then.
The Weather Channel and Weather Underground both fall under the umbrella of The Weather Company, which is now an IBM business. We deliver personalized and actionable weather data and insights to developers and businesses via APIs. We aggregate data from over 195,000 personal weather stations across the globe and combine that data with our forecasting models to provide hyper local weather forecasts. We handle 26 billion requests per day, cover 2.2 billion weather forecast locations, and reach more than 180 million consumers each month.
We are trying to map the atmosphere, not just in latitude and longitude, but also in the third dimension of elevation and over the fourth dimension (time). One major retailer, for example, uses weather data and analyzes it to manage their inventory. They know that when it’s warm out (but not too hot) and windy (but without rain), you’re more likely to buy steak. This insight comes from analyzing time series weather data and has resulted in increased sales for them.
We ingest around 40 terabytes of data per day from various sources including personal weather stations, global lightning detection systems, radar and location services, satellites, airports, and government and non-government agencies. We aggregate this data and apply forecast models to provide unique weather insights, all of which are backed by the expertise of our meteorologists and serve the data via APIs. All our consumer apps also use these APIs.
Over the years, The Weather Company has become a global weather data provider, building upon an initial Java™-based monolith infrastructure distributed via 13 data centers. At that time, we didn’t have an on-demand forecast system. Everything was a batch-based forecast system that computed the forecast, wrote it to a database and replicated it out. We had a home grown content management system that ran on The Weather Company (TWC) data centers. As a result, we were limited in terms of scaling and had a slow time to market.
We set out to fix all of that and rebooted our entire digital stack. We redesigned our big data platform and applications to run natively on the cloud. This helped us build an on-demand forecasting system instead of the batch forecasting system. We built extremely low latency (25 ms), highly available (99.99), developer-friendly APIs. The product teams within the organization use these APIs for internal development and all our mobile apps and web pages are also powered by them.
Weather is highly volatile and we make sure we provide the most precise data for every latitude-longitude combination and not just a rough estimate for a city. We can do that because of a huge network of personal weather stations distributed throughout the world combined with our forecast algorithms. To scale, we use the cloud computing service to push 4 gigabytes of data per second. The on-demand forecasting system is powered by NoSQL distribution databases like Basho’s Riak, Redis, and Cassandra which are used for distribution, replication, and data management.
By launching an API-driven platform, The Weather Company has been able to bring products and services to market faster. Our core business is creating the world’s best forecast system so that individuals and businesses can make better decisions based on weather.
Now developers can incorporate these powerful weather APIs into their apps through Bluemix®. This service lets you integrate historical and real-time weather data from The Weather Company into your IBM Bluemix application. You can retrieve weather data for an area specified by a geolocation. The data allows you to forecast, detect, and visualize disruptive weather events that might affect decision making in your application.
A cloud-based service that handles 26 billion requests a day.
The combination of technology and expertise from the two companies will enable IBM to now collect an even larger variety and higher velocity of data sets from billions of IoT sensors around the world while also serving out real-time information and insights to tens of millions of users worldwide from 2.2 billion weather forecast locations, 40 million cell phones, 50,000 flights per day, and 7x the power of the leading search engine. The Weather Company is helping clients across various industries.
Learn more about The Weather Company’s personal weather station network
- Personal Weather Station (PWS) network
- Register a PWS with Weather Underground
- Weather Underground API
- Business of Weather
Learn about IBM and The Weather Company
- IBM + The Weather Company Analytics
- Weather Means Business – Building Apps using Weather Data in Bluemix
- IBM Insights for Weather available in Bluemix
- What’s the weather like in Bluemix