It's Raining Data: How AccuWeather Drives Business Insights

The weather forecaster is harnessing its data to help companies understand how weather affects consumer behavior.
Data Analytics

It's no secret that weather plays an influential role in consumer behavior. Realizing this fact, businesses are increasingly turning to companies like AccuWeather for customized weather forecasts, data analytics, and other reporting to make more informed business decisions. In fact, the weather forecasting systems market is projected to reach $1.9 billion by 2020.

Furthermore, studies show that companies that deliver optimal customer experiences see consumers who increase their spending. Here, Rosemary Yeilding Radich, manager of business intelligence at AccuWeather, discussed growing uses of weather data and business intelligence trends with 1to1 Media Senior Writer Judith Aquino.

1to1 Media: How does AccuWeather collect data for its global weather database?

Rosemary Yeilding Radich:
We have a team at AccuWeather that's dedicated to seeking out the most high quality data sources around the globe. There's temperature, precipitation, wind speed, sunshine, and things people don't think of like humidity that we know impact consumers. We want to get as much data as we can, but it has to be data that's meaningful and high quality. We also have extensive processes to check and validate the data to make sure what we're ingesting is high quality data.

What makes the data high quality?

It has to have a high level of granularity. There's a lot of emphasis on hyper local information. People want information that doesn't just cover an entire state or zip code. They want information that's granular and specific to them. Also, when we're looking at different types of data sources, there's a big difference between data from a government weather system versus crowdsourced measurements of rainfall. When looking at crowdsourced data there are a lot of techniques we can use to aggregate the data and make sure it's high quality and give us an accurate representation of what's going on.

How quickly can AccuWeather help a marketer change an ad as the weather changes?

We have a lot of clients in the digital marketing space [for whom] we can change ads for different locations every hour or minute by minute. If they're advertising on their own website or on, we can change those ads as the forecasts change. If it's pouring at 3 p.m. but by 4 p.m. it's sunny, we can trigger the ad to change as consumer behavior changes.

Marketers also want to know whether people will go to the movies if it's raining. And at what point will consumers be purchasing their seasonal wardrobes. Every day people are making decisions that are highly impacted by the weather. One of the things we do is figure out how to make weather contextually relevant.

AccuWeather is often compared to The Weather Company, which sold many of its assets to IBM. What differentiates AccuWeather's data assets and ability to help marketers from what The Weather Company offers under IBM?

We pride ourselves on having the most accurate and granular weather data out there. We also pride ourselves on the accuracy of our forecasts. You can have the best predictive model in the world, but if you get the forecast wrong, then your model will be wrong. So say you have a model that's dependent on predicting high temperatures for different areas across the U.S. If you're basing that prediction on high temperatures and you're consistently a degree off compared to your competitors, then you're losing out no matter how good your model is. Having good data and accurate forecasts help differentiate us.

How do you see data science evolving?

One of the interesting things about data science is that it's becoming a very integrated function. We have a very diverse group of individuals. We have people in our data science teams with social science backgrounds, meteorological backgrounds and people from other areas. It's not just someone who was a database administrator or a statistician. You need to have people who are flexible and understand the data as well as the business needs. So having people who can look through the data, model it, and understand the marketing perspective, that's where the real value is in data science. If you're going to be a good data scientist, you really need to have a vast array of skills.

What are some new ways of looking at weather information that are emerging among industries?

There's a very strong link between weather and health. We have a group that looks at the relationship between weather and health and medicine. As the medical community advances, they're looking for more data points where they can start to predict things like flu outbreaks and allergies. We're also just starting to get an indication of how diseases are affected by weather since there isn't a lot of in-depth analytics yet. That's one of the big fields we'll see more work on: how weather impacts health.