What makes “The Weather Channel” the best forecaster in the world

The Weather Company, an IBM Business was named as the most accurate forecaster by ForecastWatch in 2016. ForecastWatch evaluated eleven different providers and analyzed more than 139.3 million forecasts in three different geographic regions namely the US, Europe, and Asia-Pacific. And after an extensive study, brainstorming, and analysis of the forecasts, The Weather Company was named as the market leader in this space. Today, the company's consumer brand, The Weather Channel is the most downloaded weather app globally.

So, the point is what makes The Weather Channel, the best forecast service provider in the world? Let's deep dive to discover the same.

When in 2015, IBM announced its plans to buy The Weather Company, it was still uncertain as to what this business deal was intended for. But as it turns out, this acquisition was set to transform the industries dependent on weather forecasting. With advancements in cloud and analytics technologies, weather predictions were already getting more accurate and the partnership with IBM led to integrating real-time weather insights into thousands of businesses for better decision-making.

The solutions provided by the company are helping big companies in aviation, energy, retail, insurance, media, etc. make more informed decisions and better respond to the weather impact on their businesses. This is achieved by integrating analytical models such as the SPSS Modeler by IBM with the weather data and producing actionable reports. The historical weather data can also be used to predict future impacts.

IBM's first joint product with The Weather Company was a hyperlocal weather forecast (0.2-mile to 1.2-mile resolution) called Deep Thunder to provide business clients with customized forecasts. Deep Thunder combines big data and machine-learning tools from IBM Research with The Weather Company's global forecasting model, which includes more than 195,000 personal weather stations.

The tool is aimed to help companies with critical decision-making. The data shows how changes to weather, such as temperature, affects things like consumer buying behavior, helping retailers adjust their supply chains, and manage inventories and stocks, for instance. Now, retailers cannot blame the weather for their poor financial outcomes.

Changes from the previous methods

Weather forecasting was once a subject of only atmosphere physicists and experts working on the problem. But now, applied mathematics with statistics is used in equal amounts in this field of prediction. At The Weather Company, many data scientists, machine learning experts, and Hadoop and Cassandra engineers work together for forecasting. The prediction quality increases manifold when a Ph.D. physicist and a Ph.D. applied mathematician collaborate to work on the same weather problem.

Andy Rice, Vice President, Product, and Analytics at The Weather Company attribute the accurate predictions to major improvements in the forecasting methods. Some of these include bias correction by Deep Learning methods which improve the 6-10 days forecasts, improvement in predicting electricity usage in times of weather changes, improvements in understanding weather impacts on sales of cars and houses, to name a few.

Dependence on more data

The Weather Channel feeds its forecasting engine with sources of the richest, highest-resolution data available. It not only relies on skillful numerical weather prediction models from notable agencies but also from IBM's Deep Thunder model. The engine also includes data from different sensors that collect a multitude of information as well as one of the largest troves of location data available. The richest data sets that also includes historical data are averaged — both business and consumer — for reliable and actionable weather analytics and insight.

Analytics for advanced data modeling

For creating a single forecast, 162 individual forecasts are generated and analyzed within Weather's analytical system combining a wide variety of government and private forecast models. The system combines the outcomes of machine-learning algorithms in weighing factors like temperature or precipitation on geography, time, weather type and recent forecast accuracy.

Integrating Cloud services

The analytical models have high computational costs, so The Weather Company incorporated a cloud-based platform for all types of data collected. It is reported that this platform increased data-handling capacity tenfold and handles 400 terabytes of data every day. This generates tens of millions of forecasts around the globe in microseconds.

Conclusion

The Weather Company today is ahead of the game as it has mechanism and infrastructure to make critical decisions based on data collected from the past and present which accurately observes the current conditions and gives reliable future forecasts. Consumers and businesses are making decisions on the real-time data that is subject to change every single minute. The integration of big data and analytics is proving to be the greatest development in the weather forecasting field ever and the future looks quite promising. To sum it up, all these advances will keep every person on the planet informed and help them plan and enjoy their activities accordingly.