I dread the question, but I still have to ask it: "What about external data sources?" I'm always surprised (and disappointed) by how many analytics customers don't have much of an answer. Usually, the best I hear is:
We'd love it if ____ vendor could provide that.
We're hoping to get there soon.
You can imagine my surprise, then, when GoodData customer TransTrack answered as follows:
Yeah, we're pulling weather data today. That's affecting a number of things. We're utilizing it both for ridership, and also with our electric vehicle program. Batteries perform better when it's hot, rather than cold.
We're also we're pulling in vehicle emissions data from the California Air Resources Board - they've got some really good datasets for vehicle emissions. We're also pulling demographic data, whether its ethnicity or income, that kind of stuff.
External data blended into real-time views? I don't see that often enough. But let's back up, pre-pandemic. Jackson joined TransTrack in 2018, when it was acquired by Constellation Software. Since then, TransTrack has gone from a smaller, regional company to expanding across the U.S., looking at Canada and elsewhere.
TransTrack is both a product and a consulting house. Their goal? Provide public transit companies with a comprehensive data services solution, across all aspects of their business, including Finance, HR, operations and maintenance. Jackson adds:
We pull data from all these systems. And then we do some pretty robust reporting and analytics for the CIO. And: we're totally cloud-based.
The pandemic disrupts public transit - what metrics are needed?
So when the pandemic hit, was TransTrack able to serve disrupted public transit companies? Jackson:
The thing that happened with COVID, of course, is that transportation was just crushed - especially public transit. A lot of [our clients'] ridership is at 20, 30%, and just starting to recover today... One thing that was a total culture shock to them is they always worked in offices, and now just magically, they all dispersed and they're working from home. Many of them still do today.
TransTrack's clients made the work-from-home transition - but what about their data?
Many of their systems are stovepipe systems, on-premise. So it became very difficult to get data out to everybody that needs to manage the business.
Jackson's team stepped into high gear. One goal: provide their transportation clients with metrics, across departments.
We were already consuming all that data. They're using that today to help drive their business. Since we pull business data from all across your agency, we can start pulling things together and doing compound metrics. The operations team can tell you what my on-time performance is, and how much service I ran; the finance team can tell you how much it costs and how much money I've got coming in.
But I can now tell you, by pulling all that data, what's my cost per hour? What's my cost per mile, cost per passenger, and start getting the in-depth metrics to say, "How productive are we really?"
Jackson provided sample GoodData screens his customers rely on. Here's an executive dashboard with a year-over-year view:
Executive dashboard – how is my system operating this year compared to last year?
Agencies have to report to their boards. That means providing a visual health check:
Sample board report for executive reporting
Leaning on GoodData - putting analytics to the real-time test
TransTrack has been a GoodData customer since 2017. So when the pandemic hit, TransTrack was ready:
We took the tool set, and we did two things with it. We built six or seven sets of metrics and data cubes, around different aspects of transit. But we were also providing the GoodData tool raw, as part of our product to the customers, giving them a self serve-analytics tool. So they could basically bolt on to all of our data sets.
It's like having to Tableau or Microsoft BI, but instead of having to go load all your data every time you want to do analysis, we've got all the data the agency has right there online, full time. So they've got just instant access.
Here's one useful screen provided by TransTrack: Boardings and passenger load by time of day - how many vehicles do I need based on the number of passengers I am carrying each hour? Note in the high periods, if I only had two buses an hour, with 25-70 passengers, is this too many for social distancing?
How did TransTrack know which metrics to focus on? Jackson said there's about twenty metrics most agencies measure themselves by. But the problem is: gathering that data historically isn't enough anymore.
The big problem with it is: they're trying to consume kilobytes, if not megabytes and gigabytes of information, trying to actually go compute that. Many of the current metrics, agencies were reporting two months ago, "Here's what our on-time performance was," because that was how long it takes them to assemble that data effectively and get it to the end-user.
In pandemic times, that doesn't cut it:
We're doing that in real-time. Beyond that, we've developed tons of metrics on our own as well. So we're really starting to slice and dice the data in a number of different ways. Our agencies are just like, "Wow, you can do that?" As I like to say, Our whole product is: can you do this? Or can you do that? And the answer is, "Heck yeah, we can. Let's go build it!"
Of course, during the pandemic, traffic patterns changed quickly. TransTrack's customers needed answers:
They looked at what the average loads have been on different parts of the fleets to figure out, "Do I need to add vehicles here, or can I maybe provide less service in some areas of the day?" You know, provide more effective service when it's needed more.
Pre-pandemic, it was the typical commute to work, dead time, then commute home from work. So you have those typical travel patterns. They used our tool to help analyze how those patterns have shifted. And they did shift quite a bit, so it's helping them adapt their service levels - to meet that changing travel demand.
Two more views from TransTrack customers: each one is year-over-year. It might not be pretty, but it's the data their agencies need to adapt.
Cumulative ridership by route for 2020 and 2021 - These two graphics are from two different agencies. The one on top has not recovered yet. The one on bottom has recovered somewhat, but is still being hampered by current constraints in their region.
The wrap - agency-wide data builds analytics trust
As for TransTrack, they've managed to weather the economic storm, despite their disrupted client base. To what extent does Jackson credit their data analytics savvy?
The core, obviously, is we're very sticky, and serving those core needs for the client, getting the regulatory reporting done. We have so many vendors in the industry; they create tons of data, but they just can't make it into truly useful information that the clients need.
The second aspect is one that Roman hit on [Roman Stanek, CEO of GoodData, pictured bottom left in the feature image]. A couple years ago, Roman and I were talking about the need to be providing analytics across the entire business. That was one of the reasons we ended up where we were with GoodData, because we're the same way.
I've been in this industry for twenty-something-years, and I can't tell you how many agencies I go in where planning has one set of numbers, finance has another set of numbers for the same thing. Just being able to get everybody into the same tune, where we're all agreeing on what the numbers are - it's huge. Having a tool like this, and being able to take that data agency-wide - that's really what makes us tick.
With the huge shifts in the transportation industry, from mobility to ride-sharing to self-driving cars, Jackson knows there are challenging issues ahead. But then a good data platform should adapt to new questions.