The CitySwift transit data engine is a great example of a cloud software application.
Our CEO, Brian O’Rourke spoke with TechIreland’s John O’Dea about how the business got started, and how other companies might go about creating their own cloud software. Here’s a transcript of some of the key questions and answers.
Tell us a little bit about your background and about CitySwift?
The origins of CitySwift date back many years. Myself and my co-founder Alan Farrelly have been best friends since we were at secondary school in a rural part of Longford in Ballymahon. We came to create CitySwift because my background had been in data and tech (I did information systems in college) and Alan was born into the bus industry. His family business bought their first bus back in the 1980s, and now his family home has turned into a bus depot! We left college and each did our own thing – Alan returned to the family business and I worked in technology.
One evening we were catching up with old friends, discussing the transit sector and Alan's family business. He mentioned all the data he had collected for management accounts, utilization of their vehicles... and we spotted a number of inefficiencies within the company. So we looked at the business opportunity to increase the operational efficiency of buses using data – starting with the services that Alan’s family were running.
The real change happened when we went over to the UK and we met with some senior figures in the industry. What we came to understand was that terabytes and terabytes of data were being collected by these big massive UK PLCs. But they weren't using it to generate operational efficiencies, so they turned to us and said “as an innovative young startup, is there a way that we can work together by sharing our data with you so that we can improve our net margins and improve the services we're offering to passengers.”
That was back in late 2017, early 2018. And we really started building out our platform from there – taking in massive amounts of data from all these companies, as well as open-source Smart Cities data, and then generating insights and suggestions for optimizations in a usable and actionable fashion.
"What we came to understand was that terabytes and terabytes of data were being collected by these big massive UK PLCs. But they weren't using it to generate operational efficiencies."
Is the cloud an important part of your strategy?
We've always been cloud-first and cloud-native. We knew from a very early stage that we were going to be processing large amounts of data. And we also knew that we were going to be wanting to use the latest technologies, especially in the machine learning and AI sphere.
Our strategy, from day one, was ‘how are we actually going to be able to deal with this?’. If you have one bus, you're talking gigabytes. But the transit companies we were partnering with had tens of thousands of vehicles, so we knew that we were going to be into the terabytes very, very quickly. And then we also wanted to do some really heavy and large scale AI predictions and processing on that data. The only option we had was to go cloud data from day one.
"The transit companies we were partnering with had tens of thousands of vehicles, so we knew that we were going to be into the terabytes very, very quickly."
What questions should a CEO be asking when considering cloud platforms?
Reliability and cost are very important, and also data residency. Those are the key factors – but every business is different and unique.
If you want to start to deliver value for your clients quickly, you don't really have time to develop a whole DevOps and cloud infrastructure. Our cloud provider is nearly our ‘DevOps as a Service’ and we're outsourcing a number of pieces to them.
We went with the Google Cloud Platform for some of the tools and services that they had built into it. Their data flow for pre-processing, the big data warehousing, and the AI and machine learning engines, they're baked in. The way these all seamlessly integrate and can be hooked up to each other, meant that our engineers could focus on coding, adding value rather than looking at the infrastructure underneath. It worked well for us and we're happy with how that's progressed.
"Reliability and cost are very important, and also data residency. Those are the key factors – but every business is different and unique."