Published: 08th September 2021
Real-time, accurate and hyperlocal weather predictions? Yes, please! These engineering students have done just that
Four engineering students from PES University have developed Frizzle which is now incubated at PESU Ventures. Frizzle can give you a real-time, accurate and hyperlocal weather forecast
Whenever Anantha Krishna relied on a weather app to plan his college assignments and related work, he would invariably get stuck in the rain. "There were times when my laptop and other electronic gadgets needed for experiments or projects have gotten completely destroyed due to the rain," recalls Anantha. And he wasn't the only one, this was the same case with all his other friends as well. They would head home from college only to be stuck in the rain several times. They were getting tired of inaccurate weather forecasts. That's when these four engineering students — A Anantha Krishna, Vanya Arikutharam, Abhilash V and Rishith Bhowmick from PES University got together and decided to come up with an app, Frizzle, that could provide people real-time, accurate and hyperlocal weather forecasts for any particular day.
But the team didn't just jump into designing a product. They first went around talking to people from various business industries. Anantha explains the significance, "We understood that due to bad weather forecasts, people have cancelled sporting events, entertainment events, farmers have faced loss and even restaurants have faced a lot of food wastage due to unpredictable weather change. We realised that something like this needed a long-term and more permanent solution. So, instead of limiting it to a college project, we thought of making a robust product that can help people. We also presented this idea to the PESU Ventures Labs who helped us with funds, mentorship and a lot more."
So, how exactly does their product, Frizzle, work? Anantha explains, "Frizzle has a two-part solution. We have a small portable weather station that has been designed and built by us from scratch. This weather station can be carried around to different places and it collects all the parameters from the other local weather stations that are already existing. There is also a customised algorithm that we call a hyperlocal weather engine that captures the data that we get from this portable weather station. We have built a web app that can be installed on the phone or desktop. This data from the portable station and the hyperlocal weather engine is combined to provide an accurate weather forecast."
When you asked people in the olden days and sometimes even now about the weather, they would look up to the sky and predict the weather. The same principle has been applied in the Frizzle device as well. And that's what makes their product unique. "We have installed cameras and computer vision models to check the cloud coverage and other aspects that are used as input in the hyperlocal weather engine," explains Anantha, adding, "For the most optimum coverage, we have two devices per pincode." They recently ran a pilot project using their product with the Tamil Nadu Cricket Association for the Tamil Nadu Premier League and received a positive response too. "This happened to us in June. Since then, we have been in talks with restaurants, various sports arenas and a lot more."
Regarding the challenges that they faced to bring this technology to the market, Anantha explains, "We struggled and worked hard for months because we wanted to collect hyperlocal data and getting accurate data was a challenge for us. The second challenge was to use the ground-based sensors and build a robust infrastructure for them because the sensors have to deal with heavy wind, rain and temperature. Not many people have approached weather forecasting like we have, we have fine-tuned the way we look at weather."