Startup founder Emily Blythe takes us on a training flight
She started learning to fly on the family farm when she was 12 and set up her first business while still in high school. Now, Emily Blythe aims to save travellers time, and companies billions of dollars, with better fog forecasting.
Blythe is founder and chief executive of PyperVision, which is developing a deep learning model to better predict fog and cut down on false alerts that ground aircraft and delay travellers.
In her last three years of high school, she had a company called Flat Pack, which was building a flexible toolbox for quad bikes.
“And that actually started to take off. So I did three months at uni and then thought ‘I want to do this fulltime’ and ditched uni and jumped into the business world,” the Cantabrian says after a flight on a fog-free day.
She invited the Herald as she joined her instructor for a 25-minute flight from Auckland Aero Club at Ardmore. The airport was busy with Cessnas and the Westpac Rescue Helicopter taking off and a private jet arriving. Blythe flew a Cessna 172S Skyhawk above the rich farmland and sprawling new subdivisions of south Auckland.
She’s hoping to get her PPL (private pilot’s licence) in the next few months and also hopes to grow Pyper Vision. A trip to Britain in the next few days is planned, where she’ll meet with people at London City Airport. It’s a relatively small but important London airport, close to the business hubs of Canary Wharf and the ancient city centre.
“I love creating something new. When we’re getting into fog mitigation technologies – for a long time, that’s been seen as impossible,” she says.
“And it kind of blows me away that we’ve developed so many different technologies for our aircraft over the years to combat wind and snow and all sorts and yet fog, which seems to be one of the calmest weather conditions there are, wipes everyone out.”
Fog causes losses to the sector estimated at $13 billion worldwide every year – and that’s only from the measurable commercial impact of disruption, she says.
The US National Centre for Atmospheric Research (NCAR) said fog was the second most common cause of weather-related aviation accidents, behind strong winds.
Sometimes, false fog forecasts happen, causing needless delays.
“So they’re forecasting fog, they don’t fly, and it doesn’t actually form. There’s another big problem, too,” Blythe says.
“At the moment, we’re building a forecasting product. We need to have a lot more data to better understand fog conditions at a localised level.
“And traditionally, people have used numerical weather models to forecast fog.”
Cantabrian Emily Blythe established a tech startup which enjoys backing from Icehouse and other investors.
Numerical models are the forecasting models probably most familiar to people, dividing the atmosphere into a 3D grid and using current conditions as input into mathematical models to forecast future weather.
“And the numerical weather models take a long time to process. Often they’re not very accurate at forecasting fog. The average accuracy here in New Zealand is 19%. That’s kind of unbelievable,” Blythe says.
“What we’re doing now is we’re using AI [artificial intelligence] and deep learning models, which have the ability to detect trends in fog much faster. We’re pairing them with better data on the ground – so, more spatial awareness, too.
“Our existing model at the moment is at about 32% accuracy. And once we get that additional data in there, we’re expecting to increase it quite significantly.”
She says the company’s worked with several airports and airlines.
“Hamilton Airport really kind of kickstarted us with their support and a lot of the ability to kind of learn from them at the time and understand what this would have to look like. Now we’re working with Christchurch, Dunedin and London City over in the UK as well.”
Dunedin Airport operations and infrastructure general manager Nick Rodger says it’s exciting to work with an innovative Kiwi firm using a data-driven approach to reduce the impact of fog on aviation.
“Our operational staff will work with the Pyper Vision team to use the insights gleaned from their fogcasting technology in our day-to-day airport operations, improving safety and efficiency at Dunedin Airport.”
“I pitched to him originally that I wanted to clear the fog and he thought it was impossible, but I’ve got to give him credit because he jumped straight in and has been lending a hand ever since.”
Pyper Vision’s future customers will be airlines and airports. Start-up backer Icehouse Ventures has a 37.33% stake in the company, Blythe has 39.56% and about 20 other smaller investors based in New Zealand and abroad have the rest.
Blythe says airlines need to receive better fogcasting to schedule flights, and she needs to work with airports to acquire better data on the ground.
“The model’s secure and now we need to build the infrastructure around it to deploy it.”
Blythe says she has a nimble team of five and is looking to hire a deep-learning machine engineer.
She says she would love to reach a point where Pyper Vision and its partners could build something to give people more certainty around travel plans.
For now, it’s time to prepare for her UK trip – and maybe plans for a few more flights in the Cessna to get her licence.
