comscore The Weather Channel taps into IBM's Watson and GRAF to forecast detailed and accurate weather | BGR India
News

The Weather Channel taps into IBM's Watson and GRAF to forecast detailed and accurate weather

IBM thinks that weather forecast could become accurate and detailed with the help of AI and machine learning.

Weather Channel App IBM Watson 1

Highlights

  • The Weather Channel is one of the most downloaded Weather app on Android.

  • The new app is built from the ground up with Watson intelligence and offers granular data.

  • The app offers more details about weather in an area than it did before.

Weather has always been an integral part of human society. Today, all it takes is a smartphone to know how weather conditions will vary on an hourly basis. With certain apps, you can also know the weather forecast for weeks in advance. In India, where Google‘s Android dominates the landscape, most users simply need to look at their home screen to check on the temperature for their current location or city.

Tapping on that icon will take you into the native weather application that can offer more data including the temperature in your location during the day and night. The information can even extend beyond to show how conditions are and break them into categories such as smoke, humid, et cetera. While that is the present state of weather reporting on your phone, the Weather Channel believes there is a lot more that can be done in this format. The Weather Channel, which was acquired by IBM in 2015, is now tapping into Watson, the AI Engine to power a new form of weather information, delivered right on your smartphone. Artificial intelligence and machine learning that powers it is poised to reshape a lot of industries.

While industries such as healthcare and automotive are ripe for significant transformation, regular usage like checking weather in the morning could be revolutionized too. In an interview with BGR India, Himanshu Goyal, India Business Leader, The Weather Company, said the company is now in its very early phase of undergoing such a transformation. In order to make that transformation possible, the Weather Company redesigned the app from the ground up.

“If you look back to 2017, we were not very active on app or consumer side but were more active on the business front. We had a stable app with decent following but started seeing operations of weather companies in India. In the past two years, we looked at existing weather applications and realized that consumers in India wanted an information dense application,” Goyal said in a telephonic interaction. “Based on consumer feedback, we understood that customers wanted an app that can predict the timing of rain and fluctuations. Consumers also wanted reporting on mosquito activity around time when there was a rise in viral fevers. All of this information helped us redesign our application,” he added.

The new Weather Channel app is powered by IBM Watson and uses cognitive intelligence

That experience starts when you open the app for the first time and the immediate screen shows “The Weather Channel with Watson”. Once the app opens, the first thing you do is either select a city or turn GPS ON to provide your current location. While it serves a multitude of information, the app is designed to look less technical and more click worthy. “When you look at The Weather Channel app, there is nothing technical about it but it still gets you depth of information. The app is not just built to show the change in weather conditions but help users understand its implications like surge in infections and right time to step out,” Goyal explained.

The first thing you see on the redesigned version of The Weather Channel app is weather in your city followed by a chart that breaks down the data into today, hourly and daily. The chart also depicts whether it will be sunny, windy, rainy with a symbol right below the trend chart. This is followed by the first section called “Today’s Details” which breaks down the above data into more meaningful metric such as high and low temperature for the day, sunrise and sunset time, wind speed, humidity, dew point, pressure and UV Index. It is also running an early access where users can see weather trends on an hourly basis.

The second section of the app is called “Seasonal Hub” and it offers the most critical and useful piece of information. The section offers details such as mosquito index, which indicates the weather’s influence on the activity of mosquitoes. Then there is heat index, sweat index, fog index and umbrella index depending on your location. All of these subcategories are rated as caution, high activity and in the case of umbrella index, it tells you whether you will need an umbrella. Goyal tells me that “the information is dynamic and The Weather Company is using statistical data to predict hourly weather updates with the cognitive part powered by Watson.”

The app, while mainly non-technical, has elements where data nerds can geek out and it is called Radar and Maps. The section shows radar where you see clouds, wind speed and dew point charted out. The fourth section is Videos, which is self explanatory but the app mostly show weather updates for UK, making it almost redundant. But this is also an area where the Weather Company is uniquely placed to drive adoption. During Hurricane Florence, it introduced a new technology called Green Screen Technology, which uses the green-screen immersive studio at its Atlanta headquarters, to create a mixed reality or immersive experience into weather reporting. Just imagine a meteorologist reporting about flood and showing the rise in water-level in an immersive fashion. Goyal says the technology has been tested and could soon be deployed on the app as well.

“We have started using augmented reality as a platform to show weather forecasts. These videos are created for our broadcast partners and appear on our YouTube channel as well. The initial response has been so good that we are encouraged to bring it to other platforms.”

The app which relies on vertical scrolling has the fifth section marked as Health and Activities and based on temperature around you, it offers suggestions. For instance, in Mumbai, it says the conditions are tough to play cricket and even offers health guide like whether breathing is good in your area. While all these data can be deemed critical in their individual level, The Weather Channel is stressing the most on sixth section, called Air Pollution.

The Weather Channel says the new app is completely redesigned for Android users

A report by World Health Organization said that 14 out of 15 most polluted cities in terms of PM2.5 are in India. While installing air purifiers is a good start to curb pollution indoors, there is a lot that cannot be changed in the outdoors. With the new app, the IBM-owned company forecasts Air Pollution based on PM2.5 index, which details atmospheric particulate matters having diameter less than 2.5 micrometers. The data sourced from National Air Quality Index offers insight into whether the pollution is moderate or intense. Since I started using the app, I have seen the section show air pollution hover from moderate to severe.

Clicking on this section gives more details and even offers details as to discomfort caused by prolonged exposure to polluted environment. Goyal adds that “the data is not produced by The Weather Channel and is sourced from a vendor. At times, it would be a hyperlocal information and may not be locationally accurate but corresponds to a latitude and longitude provided by the vendor.”

At this moment, the company is not contributing anything to this data. While the app can break down the information into pollutants such as NO2, SO2, PM2.5, PM10 and CO, it is limited when it comes to offering granular data. Also, the AQI shown by the app was often different from the one shown on the app from the air purifier installed at my home. This is where IBM sees major changes coming with the help of Watson and its cognitive intelligence. At CES 2019 last month, IBM introduced GRAF (Global High-Resolution Atmospheric Forecasting System), which it believes will make weather forecasting much more detailed and accurate.

Watch: How Realme phones are made

The underpinning technology relies on IBM Supercomputers and aggregates data from millions of sources. It also uses data from smartphones installed with The Weather Channel app via an opt-in option to accurately predict the weather. IBM plans to roll out GRAF later this year and detailed, accurate prediction of weather in advance could prevent disaster, prevent economic loss and save lives.

  • Published Date: February 17, 2019 2:27 PM IST