Weather forecasting has not only gotten more accurate, but the manner in which it is delivered and who it is sent to is drastically changing.
A company that is a good bellwether for this is Weather Source. This data driven weather forecasting firm recently started to use the cloud data platform Snowflake to get around the not so easy challenge of delivering petabyte levels of forecasting data to its clients, hour by hour.
The data feeding Weather Source comes from an array of organisations such as the US National Weather Service and the National Oceanic and Atmospheric Administration.
Mark Gibbas, CEO of Weather Source told Computer Business Review that they: “Pull in a tremendous amount of their information, everything from satellite information to radar to all of the weather models and weather forecasts to all of the observations that are taken at all of the weather stations and things like that.”
All of that information is then distilled down into the company’s enterprise service – On Point weather – which provides weather predictions for specific locations.
“We tend to run high resolution short term models that are designed to answer the question of what is the weather going to be in very precise increments throughout the day, kind of minute by minute weather information over a very specific area,” Gibbas notes.
The customer base for a product like this is a lot broader than you think, many people will immediately think of the agricultural or maritime sectors when it comes to the benefits of weather forecasting. Yet increasingly; savvy advertisers and service providers are also turning to services like Weather Source as Gibbas notes that: “Having knowledge of what the weather is in various parts of the country really helps them target their advertising and marketing to inspire homeowners to use their services at the right time.”
Weather Source and Snowflake
One of the challenges the company has been struggling with is how to get this data into the hands of its clients in an efficient manner that doesn’t involve them streaming tons of weather system data directly. To solve this they use the cloud data platform Snowflake.
US-based Snowflake was founded in 2012, and essentially acts as a cloud-built data warehouse that runs on Amazon S3. This month they secured £368 ($479) in a funding round that was led by Dragoneer Investment Group and Salesforce Ventures. The companies total evaluating at this point now stands at £9.5 billion ($12.4).
Rather than send data directly to a client Weather Source hosts in Snowflake and establishes data shares in the cloud. The data is uploaded either via an AWS S3 bucket or through Snowflakes own API.
Gibbas notes that through Snowflake its clients: “Get both access to the digital data for them to query directly and then they have dashboards that they can bring that into to help visualize what’s going on for whatever’s interesting to them.”
From Weather Sources point of view; Snowflake helps them reduce down the time it takes to get data directly into their clients hands as once they update the Snowflake database it is immediately pushed out to everyone they have open sharing with. Gibbas sees this as a key advantage: “Because every time we can reduce the latency on getting information to them, that means they have more time to adjust their marketing campaigns.”
Climate change is happening, and the world is slowly reorganising itself around this inevitable fact. Now organisations in sectors such as finance, insurance and infrastructure are seeking out information that can help them to understand what the climate change impact will be on projects and markets.
If expensive infrastructure, like an airports or stadiums, is going to be erected it helps to understand what the weather conditions have been for the last ten years or more in order to gleam a key insight into the weather patterns and damage that could occur 50 years out. Cities like Miami are already in the process of reshaping as they account for rising sea levels.
With their growing database this is something Weather Source is positioning itself to be in demand for as Gibbas states that: “If you want to understand what the probability of certain weather attributes happening at a particular location at a particular time of the year, we have information on that.”