Fix This: How Two Boston-Based Startups Are Forecasting Flight Delays


Image: Marcelo Cáceres

Sometimes it takes an outsider to pinpoint a pain point in an industry that’s set in its ways. In this segment of “Fix This,” we look at how two Boston-based startups, Lumo and ClimaCell, are addressing a perennial industry problem: the cost and unpredictability of flight delays.

Delays are the costly bane of airlines and travelers alike. In the US alone, the annual cost to airlines and passengers, including lost demand and indirect costs, amounted to $26.6 billion in 2017, according to the Federal Aviation Administration.

But delays and disruptions can be anticipated hours or even months before – something Boston-based Lumo is using predictive analytics to do. The underlying mechanics of Lumo’s system involve analyzing a mix of factors: time of day, day of week, airline schedules, aircraft turnaround times, historical analysis of route activities, weather, airport capacity and how different airlines handle disruptions.

Lumo already offers delay forecasting to consumers, but is currently developing a product for airlines and global distribution system companies: “My guess is that we’ll be out of both testing and prototyping midway through 2019,” Lumo co-founder Bala Chandran says.

Lumo’s simulator can predict the on-time performance of a flight at different markers before its departure.

The company’s simulator can predict the on-time performance of a flight at different markers before its departure, Chandran explains. A few months prior, it might say the schedule is looking brittle at certain connection points and that an airline might need to add extra block time. About a month before, it can say that the airline may need to activate reserve crews. “And at a week out, once the weather forecast starts to firm up, we can look at how things are going to actually play out and say we need to activate crew reserves in this location because this is going to be a hot spot three or four days from now,” he says.

“We’re also talking to an airline that’s interested in the idea of passengers getting an alert saying, ‘We think you’re going to miss your connection in two days. Here’s an alternative. Just click here and we’ll rebook you.'” Because the airline sees the delay coming before it happens, flyers no longer have to suffer the€¯aggravation of contacting a call center, waiting in line or figuring out alternatives themselves.

“ClimaCell’s engine generates forecasts by the minute using real-time observations from millions of new virtual sensing points and advanced weather modeling.” €” Itai Zlotnik, ClimaCell

Another Boston-based company operating in the delay management space is ClimaCell. The company’s HyperCast Aviation solution tracks and predicts “microweather” (what it defines as location-specific, low-altitude, short-term, actionable weather) by drawing insights from wireless networks, proprietary Internet of Things sensors and other technologies. “ClimaCell’s engine generates forecasts by the minute, in 500-meter grids using real-time observations from millions of new virtual sensing points and advanced weather modeling,” Itai Zlotnik, ClimaCell’s chief customer officer and co-founder, says.

Whereas a typical radar only senses precipitation above 1,000 feet, HyperCast detects weather from the ground up, gleaning data as to whether fog or ice is current or imminent at airports. Data garnered by HyperCast Aviation covers several time horizons: historical, real time, nowcast (0-6 hours) and forecast (6+ hours), reporting such weather parameters as precipitation intensity and type, temperature, “feels like,” wind, clouds, humidity, barometric pressure, dew point, visibility, sunrise/sunset and surface radiation – all of which can inform airline operations such as departure and landing times, deicing, and personnel preparation.

Following a successful pilot program at Boston’s Logan International Airport, JetBlue last July expanded its implementation of HyperCast Aviation across its top hubs, estimating the “potential to save the company tens of thousands of dollars per month in operational disruption costs at each of its largest airports.”

“Fix This” was originally published in the 9.1 February/March issue of APEX Experience magazine.