Why is data quality so important to the success of broadband and utility operators today?

network. Hello, and welcome to this edition of Network IQ. I'm Steve Tongish, the Chief Marketing Officer at IQGeo. And today I'm really pleased to have the Chief Technology Officer at IQGeo, David Cottingham, join us. Hi, David. Hey, Steve. Good to meet you. So today, David, the topic that we wanted to discuss is about data quality. So let me start by just making the statement that we feel that the quality of fiber network data and electric grid data is really fundamental to the success of broadband and utility operators. So the goal today is to explore that a little bit more and understand what high-quality data really means today and what are some of the opportunities with new technologies that are emerging in the future. How does that sound? Okay. So let's start with the first question. I've made the statement that high-quality data is important, but why is it important to broadband and utility operators? What's your perspective on that? You know, I think there's a really, there's a basic thing, which is, well, you should really know where your assets are, right? You should know whether you've got something underground or what side of the street it's on or wherever it is. And that, that I think intuitively makes sense to people because let's say you've got a repair needs doing. Well, where do you send the people to go and do the repair? Is it, you know, the left-hand side of the street or the right-hand side of the street? And do they have to dig or do they not? So that's kind of obvious. I think what's less obvious is when you, you think you know where things are very roughly, that's fine. But if you haven't really thought about how they're connected together from a network perspective, so if you think about kind of a network as a graph of nodes and edges, right? If you don't have those, those edges, those bits of wire in your system linked up to each other, you can't reason about, well, for example, is the shortest path between your house and mine through this set of cables on this other street or this other set of cables? And that really matters when it comes to things like fault finding. And so, you know, if we put a digger has dug through a set of fiber on one side of the street, well, who is going to be affected by that? Oh, we're not sure. We have to wait for people to call us to find out. Another one would be upgrades. How do we know? An example would be what's on a pole, right? What things are attached to a pole? How much more weight or space does it have to allow more things to be attached? Is when it comes to deploying new networks or upgrading them? Well, I need to know what's there in order to determine what else I can add without adding more poles or upgrading my poles. Another one is safety, right? So if you know more about vegetation, right? So think of in the US the past few years, there's been incidents with electrical utilities around fires and vegetation maintenance is a big deal. And so the more data you have around vegetation and where vegetation grows relative to your network, the more you can start thinking about, well, where should I go and practically cut down vegetation, for example, because it's a risk, fire risk. So there's loads of different areas that aren't immediately obvious that are super important to have really good data. Now you can start with okay data and progress to amazing data, but it's super, super important for all aspects of running a network. Yeah. Yeah. So, so you hit upon some interesting things there from, from a geospatial perspective. So from an IQ geo perspective, it's not just where these things are, but it's how they're connected. So, so, so the data is, is when we think of data, we think of intelligent data. We think of multidimensional data, multiple layers of data, don't we? Yeah. And I think when people talk about GIS is right, they tend to think about, it's sort of a map, right? You can sort of draw things on a map. And then when you digitize that and you put in a GIS, that's great. It's still a map. And it's true. You do need to know the, the precise coordinates in, in the world of, of where your, your things, your assets are. But that connectivity model of what is downstream from a particular cabinet or what is the, the number of possible homes that I could serve with this fiber backbone. So in other words, you know, I've got this fiber running down a street and how many things could be connected to it? Were I to, to convince all those customers to connect? There are lots of things where you need that, that connectivity model, which is not just about where the piece of, you know, the strand is or whatever it is. Um, that there's lots of things around. I think we'll talk about it in a minute, actually around a disaster sort of recovery and how do you guard against single points of failure? Those kinds of single points of failure detection are again, reliant on you knowing what the topology is of your network, not just where it is in, in space. Right. And you've touched on some, you're leading us into our next question, David, which, which you've already touched on, which is about the technical and commercial benefits of quality data. And you've already spoken about some of those use cases where it's clear that you won't be able to perform either commercially or technically without it. But you were, you were going to go on to talk a little more about disaster response. Yeah. And I, that for me is a really interesting one, right? Because, um, think of it today as, uh, if, if you're living in a, in a, in a map world, right, then you might give your, your disaster, the people going out to, to go and repair things in the event of a disaster. Traditionally, we've given them paper maps, right? That's what utilities have done. And they've, they've got a truck or a van and they roll out the big map on the bonnet of the truck and they figure things out. And that's okay. But again, you don't have that, that network model in mind, nor do you have necessarily up to date data. So we go back to the original point around up to date and good quality data is important. Well, that is super important when it comes to who is doing what and how are they, how is it connected to all the other things that other teams are doing? How do you make sure that when they repair something, those repairs get put back into your, your system of record because data quality isn't a one-time thing, right? You don't magically, you know, get beautiful data and then it's, you know, life is good forever after. Things change in a network all the time. So how do you make sure that, that data is kept good quality? That's really hard. But if you do it, then you get all these benefits of like quicker thought finding. So if you Steve call up your, your broadband provider and you say, you know, my connection has died. If they have really good data around how you're connected and feeds of status of all those devices in the network, they can say, aha, the reason that you're disconnected is because actually upstream from you, there's, you know, six hops away. There's something's gone wrong and I can see it, but I can see it. Yes. On the map, maps. Wonderful. I like to know where things are in space. What's now important is the fault finding of which piece of the network is at fault. And then I can send someone to the right place because yes, my data quality has to be good around where things are as well. I can send someone to the right place and they can easily find it and deal with it quickly. Not, you know, half an hour of hunting. Where's the right, you know, place? Where's the right cabinet? Oh, it's on the left hand side or the right hand side, et cetera, et cetera. We just talked about. Yeah. Yeah. And also we've, we've, we've seen how we can help customers as well with, with managing those field crews as well. So where are those field crews and where are they in proximity to, to, to the fault so that you can, you can respond quickly and, and make sure that when you say you're going to show up at two o'clock, you show up at two o'clock, all of those kinds of work. Absolutely. And even things like routing people to the right place. Right. If we know exactly where that cabinet is, not just somewhere on the street, if we can tell you, you know, you turn left, turn right, and then it's precisely 150 meters down this road. That's time saving. Massive time saving, you know, adds up over all those crews. Yeah. Yeah, absolutely. So, so this is about today and, and, and clearly I think you've articulated why data quality for these operators is so important, but I've just been, for example, at, at a big fiber show. And, and I was recently at, at Distributec, the big electrical grid show in North America. And there's a lot of conversation around machine learning, artificial intelligence. How do these new technologies fit into this picture around, around data quality? Yeah. Look, part of the, the point about data quality here is we talk about digital twins, right? And how do you, how do you create a model of the real world that you, you know, computer system that you can then use to, to reason about it? And that's where your, you know, machine learning is, is beginning. And I think will play a massive part. But it's, it starts with having good data, right? Cause garbage in, garbage out. If you give your, your magical machine learning algorithm, rubbish data, then it isn't going to give you much out. But think of it, I saw some research the other day around taking historical storm damage and looking at the network topology. And because you've got the network topology, you know, if I have a, if I break this line that will impact say 6,000 houses down downstream. Right. So I understand effectively where my single points of failure are. And if I take the storm damage, I can then work out which of those single points of failure are most at risk of, you know, being blown over or whatever it is. And so as a utility, I can then work out, well, I can underground this particular piece of the network. So I don't have to underground my entire network, just the, the, the really important single points of failure that are most likely to be damaged by the storm. And in a world where undergrounding a transmission line is something like four times the cost of putting on pylons, for example, that really, really matters. Not to mention, of course, the costs of, you know, 6,000 homes or whatever it was suddenly having no broadband or electricity or, you know, whatever it is, it's gone. That's a massive gain you can benefit from by using that machine learning type of model to simulate using a digital twin idea. What are the things I should change about my network and yet, you know, not underground the whole lot of that fast cost. Yeah. There's a bunch of other stuff around using sensors on power lines to determine when you can increase the capacity of the power line. And that's, you know, that can really make a difference because operators today use seasonal rules to work out. Well, you know, it's the weather is this, which means I can increase the capacity or decrease. And it turns out that sometimes they they over, you know, they overestimate capacity and sometimes they radically underestimate. And neither of those things is good. And so, again, having the data from those sensors and put it into the model of the network then means you can start saying, oh, well, actually, you don't need to upgrade the entire network. You just need to upgrade these pieces. And so that that kind of machine learning type of construct really, really matters. Another one would be how do you use AI in the form of image processing to help you as a network operator know have my installations or upgrades been done correctly. Now, this is more about helping you with, you know, I've got hundreds of contractors and I need to make sure they're doing the right thing. If I get them to take photos of what they're doing or what they've done, then I can then work out, ah, they've done it correctly or not. And I can immediately tell them to go fix it if they haven't. And that matters not not so much from the network model in my database, but more about making sure that I don't get future faults on the network I've just deployed. And so verifying that kind of as built is is good is super important for a yes lower maintenance costs, but also then making sure that my design that's in my GIS now matches reality. Okay. And the final one I've got, Steve, is just one on if you have in today's world, we're sort of talking about electric vehicles and so on. And then we're going to focus on how do you simulate, how do you work out how your network should change depending on, you know, you're going to buy EV tomorrow and I'm going to buy one next week. And, okay, well, suddenly we need a whole different network construct from the electricity side than we have today. What should that look like? Where do I need to go and deploy these extra, you know, resources, extra assets to enable that to happen versus, oh, the whole city needs to be done. Maybe it's by neighborhood. Maybe this neighborhood is going to is predicted to grow because I've got a data source from someone else that tells me about planning permission and so on. But I know that that network is the one to sorry, that neighborhood is the one to focus on. Right. So there are tremendous opportunities for this new technology for the optimization of, well, almost any process, the management of information over the network or the actual optimization of the network itself. But there's no shortcut in the sense that if you don't have good data quality underlying this, then you'll be compromised in being able to use that technology. So, yeah, no shortcuts. So, okay, so we've got to wrap this up, David, but I've got one last question and it needs to be a short answer. So if somebody knows they've got data quality issues, it's an issue that, you know, they want to address, they're looking to do it. Very quickly, where should they start? I think the first thing is take an honest view of where you're at. I think a lot of people assume they have better quality of data than they do. We talked a little bit about it's not a one time thing. You've got to keep, you know, what's the root of your problem? Well, if you're giving people paper and then expecting data to updates to come back, that's probably going to end badly. So use mobile solutions for GIS are clearly important. And then be prepared to go and find data sources from other people you can use to continuously update what you've got. Right. So if that's providers telling you using LIDAR where polls are, if it's address data, all those kind of things, bring that in continuously improve rather than just seeing it's a one time hit. Cool. Okay. View it in its totality. Look at it as a whole ecosystem of data. Fantastic. Okay. Well, I think we've run out of time. So thank you so much for your insights, David, and your information. And thank you for joining us with Network IQ. And we look forward to seeing you again soon. Bye-bye. Thank you. Thank you.