Analysing the data team behind life-saving drone flights

Keller Rinaudo, the CEO of fast-growing drone logistics company, Zipline could not have said it better:

“All our customers really care about is whether something goes from point A to point B fast enough to save someone’s life. We could be using magical dragons for all they care; just as long as the job gets done.”

And as well they should too; because why should they care about anything else, except whether a product they ordered will get to where it needed exactly when it is needed? Of course, in their spare time, people might marvel at the fact that the chain reaction of events that happen from the moment they order a medical product to the moment it is catapulted off the runway and on its way to hospital takes roughly five minutes.

It is a really cool process. And watching it live in action might be fun too for the first few times.

But did you know that behind all the over one million successful flights Zipline has so far completed in Ghana, Rwanda and the USA is a data analytics team that works on each byte of data sent by Zips after every flight, to check how it went on and how it can be improved?

At Zipline, that thankless behind-the-scenes service is provided by Databricks, a San Francisco, California based data analytics company that helps organisations prepare their data for analytics and empowers the adoption of data science and data-driven decisions among organisations.

“While drones are the vehicle for this global medical delivery service, the unsung heroes are the data behind the deliveries that makes them safe, reliable, and scalable,” said Databricks in a recent statement about its work for Zipline. “Getting that job done successfully requires the ability to ingest and analyse huge chunks of time-series data in real time. This data is produced every time a drone takes flight, and includes performance data, in-flight battery management, regional weather patterns, geographic obstacles, landing errors and a litany of other information that has to be processed.

“Every Zipline flight generates a gigabyte of data with potential life-or-death consequences, especially if it throws a Zipline drone (the Zip) off course. But accessing and federating the data for both internal and external decision making was easier said than done before Databricks, as Zipline didn’t have an efficient way of harnessing and sharing the data across the organisation and their supply chain partners.”

The Zipline data department realised it was short a centralised data platform where they could not only allow them to support their real-time data needs, but also allow for access to data and insights by various teams, so they would make smarter decisions that could save more lives in real time.

Socially distanced. The Zipline data team

They turned to Databricks.

The company says that with its platform, Zipline’s data team and their colleagues throughout the company (wherever they are) are able to access all of the information they need to accurately measure success, find the metrics that relate to customer experiences or logistics, and improve on them exponentially as more data is ingested and machine learning models are refined.

“Within a data-driven culture, it’s important to ensure everyone is able to make decisions with the right data,” said Matt Fay; data team lead at Zipline. “Databricks not only helps our data team, but also everyone in the company deliver insights faster. It also empowers us to not only build the tools and utilities necessary to access our data, but others in the organisation can easily build visualisations and start extracting actionable insights from the data to solve pressing issues.”

With real-time access to thousands of flights’ worth of data, Zipline’s engineers can quickly write custom analyses in order to figure out what’s going on in the field, determine if there are any issues occurring, their origins; and how those various factors could impact operations. Such information will make it easier for the team to decide on whether to replace or retrieve certain drones — and how each decision affects the healthcare facilities they serve.

Diagnosing real-world problems in real time is a welcome advance in healthcare, which gained an especially crucial responsibility when the COVID-19 pandemic started spreading around the world. With medical drone deliveries also rising to prominence during the same time, their importance as an effective tool to fight the pandemic and help set the world back to normal, everyday problems that occupied it before December 2019 could not be more emphasised.

“We’re working here because we’re saving lives,” says Whitney Huang, Zipline’s Embedded Software Engineer. “We’re diagnosing real-world problems in real time, finding anomalies as they happen, and fine-tuning our algorithms so that we become more successful with every flight.

“Databricks gives us confidence in our operations, and enables us to continuously improve our technology, expand our impact, and provide lifesaving aid where and when it’s needed, every single day.”

It is food for thought for Africans thinking of a career in the drone industry – it is not only in drone piloting, drone building, drone flight instructing, or drone maintenance where opportunities lie. You can analyse drone flight data too. We might never sing your name, but you will be an essential worker in the industry. Before coronavirus, we took the job that doctors, nurses, emergency services technicians and their fellow colleagues in the healthcare industry did for granted; after all, all they did was save our lives.

Now we cannot sing their praises enough.


Leave a Comment


Welcome! Login in to your account

Remember me Lost your password?

Lost Password