top of page

E41: How is AI powering environmental monitoring in major cities?

Updated: Feb 5, 2023

Catch this Team Check-in with Greg Johnston, President of Carl Data Solutions, the Vancouver-based predictive analytics startup that recently signed an environmental monitoring deal with Los Angeles Sanitation Districts, a utility that serves over 5.6 million people.

Click here to subscribe to our podcast and newsletter. Our special thanks go to Microsoft Canada and Cinchy, the Dataware platform for sponsoring the 4th series of the Ask AI podcast.

Vancouver-based Carl Data Solutions got their start in real-time environmental analytics with FlowWorks in 2014. Since then, they have picked up 70+ water treatment customers across North America including the cities of Boston, Miami, Dallas, Seattle, and Toronto.

In 2019, the company was awarded a leadership role in the Fresh Water Data Commons Project consortium at Canada's Digital Supercluster and the IP developed in the course of that project led to a new generation of products which began launching in 2022.

Most recently, their AI technology helped to close the company's largest-ever sale to Los Angeles Sanitation Districts, a utility company that serves over 5.6 million people. It also attracted a new strategic partner called K2 Geospatial who are focused in smart city, seaport, airport and other infrastructure utilities across North America and Europe.

Stream the video or download the podcast for this episode as Ask AI host Carolyne Pelletier checks in with Greg Johnston, President of Carl Data Solutions, the Vancouver-based predictive analytics startup to ask the question:

How is artificial intelligence supporting environmental monitoring in major cities?

Download episode

Support episode

Like, comment, share, or tag folks and help this episode get noticed 👍

Stream episode

Click below to play the YouTube video. Don't forget to like and subscribe!

Talk to our bot

This episode and hundreds of AI resources are available from the Ask AI chatbot.

Special thanks to our chatbot partner Ada for providing this World-class made-in-Canada technology!

Episode links

Related episodes

Episode credits

About Ask AI

Ask AI is a volunteer project exploring Canada's World-class artificial intelligence sector. Our projects include thought leader interviews, team check-ins, and live events. To learn more about sponsorships, guest blogging, or volunteering please contact us or ask our chatbot.

Episode transcript

Please note: this transcript was generated by an artificial intelligence and some typos are invevitable:

Carolyne: so Greg I'd love to know more about yourself, your background and more about Carl Data Solutions.

Greg: Sure. Yeah. So I've been with Carl since 2015. My background is in a database structures.

Statistics are mostly in database marketing actually. With Carl though we're focused on in more environmental things. Carl's based in Vancouver. Company specializes in AI driven predictive analytics. We're a publicly traded company and took her single is a CRL on the CSC.

And we're also cross-listed him a small cap stock exchanges in the U S and Germany.

Carolyne: great. Thank you so much. And can you tell me more about how Carl was.

Greg: Sure. Carl Carl started back in 2015. We started with some technology but the the big starting point for us was the purchase of an application or software as a service application called FlowWorks for.

So now has about 70 different clients across North America dealing with stormwater system utilities and actually really serves as an application that governs areas where there's tens of billions of people.

Carolyne: And I'd love to know more about the name Carl Data Solutions.

How did that all come about?

Greg: Well actually the name Carl was bad named for the company. We're going to come up with the versions of why the name exists. Carl became numerous after that the company actually started doing it. My favorite one was it has to do with Carl Young.

A theory of the interconnectedness of all things. So we thought that was pretty appropriate for us since we're trying to connect basically everything. But really yeah it's a project name that turned into a company name as well to meetings.

Carolyne: Wonderful Greg. And can you tell me what's new at Carl?

Greg: Sure. We just finished a big digital supercluster project. It was sponsored by the federal government. We had partners including tech resources Microsoft of living lakes Canada university of Victoria in genome BC all involved in the project went on for about 24 months.

We got it to. 2019. So check it was a significant undertaking a big amount of work that we built and we focused on constructing an anti-cancer environmental monitoring solution and selected a town in Southeastern BC called Nelson and built a water balance model and a bunch of other.

Proof of concept so neat tools that came out

Carolyne: of it. Great. And can you tell me more about this project? What is it called and what was the goal of this project that came out of this clever.

Greg: The the the platform we built we felt we call it flow H2O and it's a comprehensive end-to-end environmental data monitoring system.

It's the great part about it is it's incredibly scalable. So we. The idea was to add many many locations and we still hope to do that. So we are in addition to the existing clients where we've kind of rolled them into the new technology and now we're roughing out that technology for new markets and new customers.

The idea is we can handle a massive amount of information coming in from multiple disparate places and be able to provide the analytics and the real-time streaming analytics that people need to actually monitor and and make predictions on on infrastructure and environmental conditions and events.

Carolyne: Wonderful. And I'd love to hear more about the technology behind you know what type of AI powers this platform?

Greg: Sure. So one of the greatest pieces of technology that came out of the project was a. Scripting engine that allows a year to create new channels of data by using machine learning feedback loops these kinds of things and apply it directly to real-time data.

That's coming into the system. So as the data comes in it's alongside that reported data it's creating these new calculated channels based on these pretty sophisticated algorithms and with that. And some of some work you can come up with some some great predictions.

And of course the great part about machine learning type systems is they they get better with age with more data. They get more time and more feedback you provide. The system just gets better and better and better. So it's a pretty powerful feature and it says centers prominently in a lot of the products that we're coming out with now.

Carolyne: Great. And so I'd love to hear more about the platform itself. Somehow is the data presented. How does it help your users kind of make better decisions with these predictions?

Greg: You can access the data from the platform through an API service or one of our applications. So they have several applications that actually work based on the backend that we've built in those applications.

And typically across the board people will bring the information into GIS systems or graphing reporting software. You can do lots of different things. It's like a Swiss army knife for real time streaming data. You can do all sorts of great. Things with it including it are alarming. So if a particular threshold is met based on four or five different variables for a few different sites or something new is discovered based on an algorithm that's running against the data people, the correct people can be notified.

This is incredibly beneficial whether you're using applications or. We're merging it with something that's that like a GIS system that's say a mining company for instance is working but there's lots of difference. We we made it very functional and an easy to integrate to into different systems.

Carolyne: Wonderful. And can you talk about any real world examples of how the system uses any proof of concepts that you'd like to mention?

Greg: Yeah the proof of concept for the data commons was a a couple of predictive channels streams of data.

That the use a a water balance model for watershed near there Nelson what it does is it creates the predictions of future outflows of water from that particular watershed. So you could see based on all the different variables that are aggregated together and the algorithms and the updates to the underlying water balance model.

What the prediction will be saying seven days from now. So it's a great tool for protecting infrastructure if we're looking at what to expect a mining company or a municipality needs to know of course and especially this is a. Relevant given the flooding that we had in the fall, what to expect in our upcoming system that's going to come.

That's going to come through what can we expect that that outflow to be or and how can we and how is it going to impact our critical infrastructure?

Carolyne: Absolutely. rightful. And can you talk a bit more about what are the benefits of your AI platform?

Greg: Sure. Using our platform and the applications. So the API interfaces that we have allows you to manipulate the streaming data that's being collected from sensors in the field and aggregate all these different data sources together in real time.

And then. The algorithms to get the predictive results like my infrastructure are going to wash her way. And if there's a big environmental vendor how big. A storm in terms of the amount of precipitation camera system that they accommodate without needing a massive capital investment to in an upgrade.

And it is the system itself degrading over time due to age and cracks and things like that that are allowing it to become less efficient as it was when it was first developed.

Carolyne: Related to machine learning and AI, what are the main benefits that are coming onto your platform using these technologies?

Greg: I think that the biggest benefit of Raleigh is time savings and the ability to do things in the amount of time that makes sense that its actions can actually be addressed and and tandem to. Yeah for instance with you when you are applying machine learning it's looking for a pattern recognition.

And so you know that saves a ton of time for some mundane tasks that need to be performed and you have your data. Just being you know multiplying year after year you're getting more and more and more data sampling at the increased frequency. You're going to run on steam. I run out of people to be able to accommodate and work with that data.

What our platform does is it provides the tools to be able to harness all that information so you can become scalable and be able to add new sources without degradation in your operations at all.

Carolyne: Great. And can you talk a bit about some of your clients and any you know recent raises that you've done to.

Greg: Sure sure. We just signed actually our biggest client today which is LA county. Nice. We have Los Angeles as it is the city is a client but now we've expanded and we have LA county which is huge.

It's over 5 million people it's encompasses about 78 cities and love. 14 or 1500 miles of sewer systems. And so there's just a lot of infrastructure there that that they're we're monitoring. We've got other partnerships now we're working together to expand our platform and merge it with other technologies in particular partnership with the Montreal based firm called K2 geospatial.

They sell a GIS mapping integration platform LJ map 500. And we can integrate our system with theirs to provide a much bigger bang for the buck for both of our client lists. And so it's going to be very exciting to get into this part of it. That's great. We were also able to strengthen the clients that we have the the products that were developed from the digital supercluster and the direction that we were going in we were able to bring in a new CEO John Charles spinoff.

Liz I came in and helped us raise significantly a lot of capital but that really puts us on a firm footing moving forward and gives us both leadership and then the capital that we need to really take things to the next level.

Carolyne: wonderful. And Greg can you tell us about what's next that Carl Data Solutions

Greg: you have a number of. Solutions that are there coming out to the that are all based on a lot of the work we did with the digital Cipro cluster project the one that's the eminent and the that it's that I'm really excited about is something we're calling auto QA QC.

It's a quality assurance. Review control basically what this does is use machine learning to pick up the patterns and dances that their sites look for for bias and then eliminate that bias or at least reduce it. So you're not if you have a lot of calculations that are driving reports and Walter arms and things like that.

It's not based on removing as much bias from the data beforehand as possible. This is a massive time saver for companies because typically you know that that's 90% of the data the data project is making sure that you've got a good underlying data set. So this is really helpful.

Carolyne: wonderful. And you can talk about a bit more about the type of bias that you're seeing and maybe some more definitions surrounding this term.

Greg: Sure, bias really really needs all the leaves in the data. So data points that are showing up that shouldn't be there and then we can prove that they shouldn't be there.

And that's what we use machine learning for. If there's something that happens in the fields like voltage spikes and calibrations or I go off and lead to. Bad bad readings that come in and or misreadings and you know a lot of different things. All those things can factor in.

And if they're significant enough it can impact your end calculations of the formulas and the algorithms that are running a guest to create these reports and alarms and then whatnot. So scrubbing that data I would say to eliminate the anomalies is. It's extremely important. It's a massive time saver.

Carolyne: Yeah absolutely. All right. Wonderful Greg. And if people wanted to get more information about Carl Data Solutions where could they go and what could they do?

Greg: Sure. They can reach us or they can find more information on a website or specifically for the FlowWorks application at

There we'll find lots of new and relevant information.

Carolyne: Great. Thank you so much for joining us

Greg: today. Thank you.

206 views0 comments


bottom of page